There isn't an explicit stolen memory base register on gen2.
Some old comment in the i915 code suggests we should get it via
max_low_pfn_mapped, but that's clearly a bad idea on my MGM.
The e820 map in said machine looks like this:
BIOS-e820: [mem 0x0000000000000000-0x000000000009f7ff] usable
BIOS-e820: [mem 0x000000000009f800-0x000000000009ffff] reserved
BIOS-e820: [mem 0x00000000000ce000-0x00000000000cffff] reserved
BIOS-e820: [mem 0x00000000000dc000-0x00000000000fffff] reserved
BIOS-e820: [mem 0x0000000000100000-0x000000001f6effff] usable
BIOS-e820: [mem 0x000000001f6f0000-0x000000001f6f7fff] ACPI data
BIOS-e820: [mem 0x000000001f6f8000-0x000000001f6fffff] ACPI NVS
BIOS-e820: [mem 0x000000001f700000-0x000000001fffffff] reserved
BIOS-e820: [mem 0x00000000fec10000-0x00000000fec1ffff] reserved
BIOS-e820: [mem 0x00000000ffb00000-0x00000000ffbfffff] reserved
BIOS-e820: [mem 0x00000000fff00000-0x00000000ffffffff] reserved
That makes max_low_pfn_mapped = 1f6f0000, so assuming our stolen
memory would start there would place it on top of some ACPI
memory regions. So not a good idea as already stated.
The 9MB region after the ACPI regions at 0x1f700000 however
looks promising given that the macine reports the stolen memory
size to be 8MB. Looking at the PGTBL_CTL register, the GTT
entries are at offset 0x1fee00000, and given that the GTT
entries occupy 128KB, it looks like the stolen memory could
start at 0x1f700000 and the GTT entries would occupy the last
128KB of the stolen memory.
After some more digging through chipset documentation, I've
determined the BIOS first allocates space for something called
TSEG (something to do with SMM) from the top of memory, and then
it allocates the graphics stolen memory below that. Accordind to
the chipset documentation TSEG has a fixed size of 1MB on 855.
So that explains the top 1MB in the e820 region. And it also
confirms that the GTT entries are in fact at the end of the the
stolen memory region.
Derive the stolen memory base address on gen2 the same as the
BIOS does (TOM-TSEG_SIZE-stolen_size). There are a few
differences between the registers on various gen2 chipsets, so a
few different codepaths are required.
865G is again bit more special since it seems to support enough
memory to hit 4GB address space issues. This means the PCI
allocations will also affect the location of the stolen memory.
Fortunately there appears to be the TOUD register which may give
us the correct answer directly. But the chipset docs are a bit
unclear, so I'm not 100% sure that the graphics stolen memory is
always the last thing the BIOS steals. Someone would need to
verify it on a real system.
I tested this on the my 830 and 855 machines, and so far
everything looks peachy.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Link: http://lkml.kernel.org/r/1391628540-23072-3-git-send-email-ville.syrjala@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
For gen2 devices we're going to need another way to determine
the stolen memory base address. Make that into a vfunc as well.
Also drop the bogus inline keyword from gen8_stolen_size().
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Link: http://lkml.kernel.org/r/1391628540-23072-2-git-send-email-ville.syrjala@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull x86 fixes from Peter Anvin:
"Quite a varied little collection of fixes. Most of them are
relatively small or isolated; the biggest one is Mel Gorman's fixes
for TLB range flushing.
A couple of AMD-related fixes (including not crashing when given an
invalid microcode image) and fix a crash when compiled with gcov"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, microcode, AMD: Unify valid container checks
x86, hweight: Fix BUG when booting with CONFIG_GCOV_PROFILE_ALL=y
x86/efi: Allow mapping BGRT on x86-32
x86: Fix the initialization of physnode_map
x86, cpu hotplug: Fix stack frame warning in check_irq_vectors_for_cpu_disable()
x86/intel/mid: Fix X86_INTEL_MID dependencies
arch/x86/mm/srat: Skip NUMA_NO_NODE while parsing SLIT
mm, x86: Revisit tlb_flushall_shift tuning for page flushes except on IvyBridge
x86: mm: change tlb_flushall_shift for IvyBridge
x86/mm: Eliminate redundant page table walk during TLB range flushing
x86/mm: Clean up inconsistencies when flushing TLB ranges
mm, x86: Account for TLB flushes only when debugging
x86/AMD/NB: Fix amd_set_subcaches() parameter type
x86/quirks: Add workaround for AMD F16h Erratum792
x86, doc, kconfig: Fix dud URL for Microcode data
The following path will cause array out of bound.
memblock_add_region() will always set nid in memblock.reserved to
MAX_NUMNODES. In numa_register_memblks(), after we set all nid to
correct valus in memblock.reserved, we called setup_node_data(), and
used memblock_alloc_nid() to allocate memory, with nid set to
MAX_NUMNODES.
The nodemask_t type can be seen as a bit array. And the index is 0 ~
MAX_NUMNODES-1.
After that, when we call node_set() in numa_clear_kernel_node_hotplug(),
the nodemask_t got an index of value MAX_NUMNODES, which is out of [0 ~
MAX_NUMNODES-1].
See below:
numa_init()
|---> numa_register_memblks()
| |---> memblock_set_node(memory) set correct nid in memblock.memory
| |---> memblock_set_node(reserved) set correct nid in memblock.reserved
| |......
| |---> setup_node_data()
| |---> memblock_alloc_nid() here, nid is set to MAX_NUMNODES (1024)
|......
|---> numa_clear_kernel_node_hotplug()
|---> node_set() here, we have an index 1024, and overflowed
This patch moves nid setting to numa_clear_kernel_node_hotplug() to fix
this problem.
Reported-by: Dave Jones <davej@redhat.com>
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Tested-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
Reported-by: Dave Jones <davej@redhat.com>
Cc: David Rientjes <rientjes@google.com>
Tested-by: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
On-stack variable numa_kernel_nodes in numa_clear_kernel_node_hotplug()
was not initialized. So we need to initialize it.
[akpm@linux-foundation.org: use NODE_MASK_NONE, per David]
Signed-off-by: Tang Chen <tangchen@cn.fujitsu.com>
Tested-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
Reported-by: Dave Jones <davej@redhat.com>
Reported-by: David Rientjes <rientjes@google.com>
Tested-by: Dave Jones <davej@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
For additional coverage, BorisO and friends unknowlingly did swap AMD
microcode with Intel microcode blobs in order to see what happens. What
did happen on 32-bit was
[ 5.722656] BUG: unable to handle kernel paging request at be3a6008
[ 5.722693] IP: [<c106d6b4>] load_microcode_amd+0x24/0x3f0
[ 5.722716] *pdpt = 0000000000000000 *pde = 0000000000000000
because there was a valid initrd there but without valid microcode in it
and the container check happened *after* the relocated ramdisk handling
on 32-bit, which was clearly wrong.
While at it, take care of the ramdisk relocation on both 32- and 64-bit
as it is done on both. Also, comment what we're doing because this code
is a bit tricky.
Reported-and-tested-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: http://lkml.kernel.org/r/1391460104-7261-1-git-send-email-bp@alien8.de
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
- Revert "xen/grant-table: Avoid m2p_override during mapping" as it broke Xen ARM build.
- Fix CR4 not being set on AP processors in Xen PVH mode.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJS8AyQAAoJEFjIrFwIi8fJbD4IAJssMuaLI5CRsSWBgDFHHDFt
srVJpDOYQiDr/TxkwFCVcL4sFy9Htb3KMArU4eIBl6uMqQbGa+3rHyXcHYI219YY
XH3D8RG+9JChwsxtaeUEzwx1C8ehcygD34vtdcoQXa7eBuEi4TL3HeLifR+HrXKO
UdFrTA34FmvpVFbSuRXkZh5sd6ca9et9xHuQHM8SIY6pVokY6xaEYOp17tfPZpwM
7A6LFjUjXeugHC2L3+/H8UOHA9nSZQvnMiZOWq2Cusc2Dt2V7emzgk2wcc2CHttf
EA6GbtiJzHqMPmt5EjubI9hHdSMB31HpY4hnQE38+ucl+BwiSdRE9z2Rm4TYClg=
=IX4M
-----END PGP SIGNATURE-----
Merge tag 'stable/for-linus-3.14-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull Xen fixes from Konrad Rzeszutek Wilk:
"Bug-fixes:
- Revert "xen/grant-table: Avoid m2p_override during mapping" as it
broke Xen ARM build.
- Fix CR4 not being set on AP processors in Xen PVH mode"
* tag 'stable/for-linus-3.14-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen/pvh: set CR4 flags for APs
Revert "xen/grant-table: Avoid m2p_override during mapping"
CONFIG_X86_32 doesn't map the boot services regions into the EFI memory
map (see commit 700870119f ("x86, efi: Don't map Boot Services on
i386")), and so efi_lookup_mapped_addr() will fail to return a valid
address. Executing the ioremap() path in efi_bgrt_init() causes the
following warning on x86-32 because we're trying to ioremap() RAM,
WARNING: CPU: 0 PID: 0 at arch/x86/mm/ioremap.c:102 __ioremap_caller+0x2ad/0x2c0()
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.13.0-0.rc5.git0.1.2.fc21.i686 #1
Hardware name: DellInc. Venue 8 Pro 5830/09RP78, BIOS A02 10/17/2013
00000000 00000000 c0c0df08 c09a5196 00000000 c0c0df38 c0448c1e c0b41310
00000000 00000000 c0b37bc1 00000066 c043bbfd c043bbfd 00e7dfe0 00073eff
00073eff c0c0df48 c0448ce2 00000009 00000000 c0c0df9c c043bbfd 00078d88
Call Trace:
[<c09a5196>] dump_stack+0x41/0x52
[<c0448c1e>] warn_slowpath_common+0x7e/0xa0
[<c043bbfd>] ? __ioremap_caller+0x2ad/0x2c0
[<c043bbfd>] ? __ioremap_caller+0x2ad/0x2c0
[<c0448ce2>] warn_slowpath_null+0x22/0x30
[<c043bbfd>] __ioremap_caller+0x2ad/0x2c0
[<c0718f92>] ? acpi_tb_verify_table+0x1c/0x43
[<c0719c78>] ? acpi_get_table_with_size+0x63/0xb5
[<c087cd5e>] ? efi_lookup_mapped_addr+0xe/0xf0
[<c043bc2b>] ioremap_nocache+0x1b/0x20
[<c0cb01c8>] ? efi_bgrt_init+0x83/0x10c
[<c0cb01c8>] efi_bgrt_init+0x83/0x10c
[<c0cafd82>] efi_late_init+0x8/0xa
[<c0c9bab2>] start_kernel+0x3ae/0x3c3
[<c0c9b53b>] ? repair_env_string+0x51/0x51
[<c0c9b378>] i386_start_kernel+0x12e/0x131
Switch to using early_memremap(), which won't trigger this warning, and
has the added benefit of more accurately conveying what we're trying to
do - map a chunk of memory.
This patch addresses the following bug report,
https://bugzilla.kernel.org/show_bug.cgi?id=67911
Reported-by: Adam Williamson <awilliam@redhat.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Matthew Garrett <mjg59@srcf.ucam.org>
Cc: Rafael J. Wysocki <rjw@rjwysocki.net>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
It can take some time to validate the image, make sure
{allyes|allmod}config doesn't enable it.
I'd say randconfig will cover it often enough, and the failure is also
borderline build coverage related: you cannot really make the decoder
test fail via source level changes, only with changes in the build
environment, so I agree with Andi that we can disable this one too.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Paul Gortmaker paul.gortmaker@windriver.com>
Suggested-and-acked-by: Andi Kleen andi@firstfloor.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
During bootup in the 'probe_page_size_mask' these CR4 flags are
set in there. But for AP processors they are not set as we do not
use 'secondary_startup_64' which the baremetal kernels uses.
Instead do it in this function which we use in Xen PVH during our
startup for AP processors.
As such fix it up to make sure we have that flag set.
Signed-off-by: Mukesh Rathor <mukesh.rathor@oracle.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
This reverts commit 08ece5bb23.
As it breaks ARM builds and needs more attention
on the ARM side.
Acked-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
With DISCONTIGMEM, the mapping between a pfn and its owning node is
initialized using data provided by the BIOS. However, the initialization
may fail if the extents are not aligned to section boundary (64M).
The symptom of this bug is an early boot failure in pfn_to_page(),
as it tries to access NODE_DATA(__nid) using index from an unitialized
element of the physnode_map[] array.
While the bug is always present, it is more likely to be hit in kdump
kernels on large machines, because:
1. The memory map for a kdump kernel is specified as exactmap, and
exactmap is more likely to be unaligned.
2. Large reservations are more likely to span across a 64M boundary.
[ hpa: fixed incorrect use of "pfn" instead of "start" ]
Signed-off-by: Petr Tesarik <ptesarik@suse.cz>
Link: http://lkml.kernel.org/r/20140201133019.32e56f86@hananiah.suse.cz
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Pull core debug changes from Ingo Molnar:
"This contains mostly kernel debugging related updates:
- make hung_task detection more configurable to distros
- add final bits for x86 UV NMI debugging, with related KGDB changes
- update the mailing-list of MAINTAINERS entries I'm involved with"
* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
hung_task: Display every hung task warning
sysctl: Add neg_one as a standard constraint
x86/uv/nmi, kgdb/kdb: Fix UV NMI handler when KDB not configured
x86/uv/nmi: Fix Sparse warnings
kgdb/kdb: Fix no KDB config problem
MAINTAINERS: Restore "L: linux-kernel@vger.kernel.org" entries
- Xen ARM couldn't use the new FIFO events
- Xen ARM couldn't use the SWIOTLB if compiled as 32-bit with 64-bit PCIe devices.
- Grant table were doing needless M2P operations.
- Ratchet down the self-balloon code so it won't OOM.
- Fix misplaced kfree in Xen PVH error code paths.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJS68IQAAoJEFjIrFwIi8fJAWgH/j4HStEey3rgGcqwIWSHkkap
+t55wsrT8Ylq6CzZjaUtCo3pB7HotW526x/0rA2pxVqHn/8oCN/1EtdrNtYm/umX
qOoda+db5NIjAEGVLWSLqGyokJQDrX/brXIWfYR300e9fnJi7yT/rFC4QHoZVUYl
5LME8XH/jE012vvYelNu6DbbodlRmVCT8hctJS+eB5ER2WmtD9Pkw4GybEXPVYJz
hE0Ts1DN91nKP2FGJb+mfB9UFT5X8i00akAK+Qc1R3sRnRh6eRoNV8dgyCnudKpO
UPEdiAZvgij+mzlgIYSz6nKH0U/VbvRsG3lc3i5Si3o+vR3CYPCkvzOGX2d0rjw=
=7cxW
-----END PGP SIGNATURE-----
Merge tag 'stable/for-linus-3.14-rc0-late-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull Xen bugfixes from Konrad Rzeszutek Wilk:
"Bug-fixes for the new features that were added during this cycle.
There are also two fixes for long-standing issues for which we have a
solution: grant-table operations extra work that was not needed
causing performance issues and the self balloon code was too
aggressive causing OOMs.
Details:
- Xen ARM couldn't use the new FIFO events
- Xen ARM couldn't use the SWIOTLB if compiled as 32-bit with 64-bit PCIe devices.
- Grant table were doing needless M2P operations.
- Ratchet down the self-balloon code so it won't OOM.
- Fix misplaced kfree in Xen PVH error code paths"
* tag 'stable/for-linus-3.14-rc0-late-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen/pvh: Fix misplaced kfree from xlated_setup_gnttab_pages
drivers: xen: deaggressive selfballoon driver
xen/grant-table: Avoid m2p_override during mapping
xen/gnttab: Use phys_addr_t to describe the grant frame base address
xen: swiotlb: handle sizeof(dma_addr_t) != sizeof(phys_addr_t)
arm/xen: Initialize event channels earlier
two s390 guest features that need some handling in the host,
and all the PPC changes. The PPC changes include support for
little-endian guests and enablement for new POWER8 features.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)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=yd0Z
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull more KVM updates from Paolo Bonzini:
"Second batch of KVM updates. Some minor x86 fixes, two s390 guest
features that need some handling in the host, and all the PPC changes.
The PPC changes include support for little-endian guests and
enablement for new POWER8 features"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (45 commits)
x86, kvm: correctly access the KVM_CPUID_FEATURES leaf at 0x40000101
x86, kvm: cache the base of the KVM cpuid leaves
kvm: x86: move KVM_CAP_HYPERV_TIME outside #ifdef
KVM: PPC: Book3S PR: Cope with doorbell interrupts
KVM: PPC: Book3S HV: Add software abort codes for transactional memory
KVM: PPC: Book3S HV: Add new state for transactional memory
powerpc/Kconfig: Make TM select VSX and VMX
KVM: PPC: Book3S HV: Basic little-endian guest support
KVM: PPC: Book3S HV: Add support for DABRX register on POWER7
KVM: PPC: Book3S HV: Prepare for host using hypervisor doorbells
KVM: PPC: Book3S HV: Handle new LPCR bits on POWER8
KVM: PPC: Book3S HV: Handle guest using doorbells for IPIs
KVM: PPC: Book3S HV: Consolidate code that checks reason for wake from nap
KVM: PPC: Book3S HV: Implement architecture compatibility modes for POWER8
KVM: PPC: Book3S HV: Add handler for HV facility unavailable
KVM: PPC: Book3S HV: Flush the correct number of TLB sets on POWER8
KVM: PPC: Book3S HV: Context-switch new POWER8 SPRs
KVM: PPC: Book3S HV: Align physical and virtual CPU thread numbers
KVM: PPC: Book3S HV: Don't set DABR on POWER8
kvm/ppc: IRQ disabling cleanup
...
Passing a freed 'pages' to free_xenballooned_pages will end badly
on kernels with slub debug enabled.
This looks out of place between the rc assign and the check, but
we do want to kfree pages regardless of which path we take.
Signed-off-by: Dave Jones <davej@fedoraproject.org>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
The grant mapping API does m2p_override unnecessarily: only gntdev needs it,
for blkback and future netback patches it just cause a lock contention, as
those pages never go to userspace. Therefore this series does the following:
- the original functions were renamed to __gnttab_[un]map_refs, with a new
parameter m2p_override
- based on m2p_override either they follow the original behaviour, or just set
the private flag and call set_phys_to_machine
- gnttab_[un]map_refs are now a wrapper to call __gnttab_[un]map_refs with
m2p_override false
- a new function gnttab_[un]map_refs_userspace provides the old behaviour
It also removes a stray space from page.h and change ret to 0 if
XENFEAT_auto_translated_physmap, as that is the only possible return value
there.
v2:
- move the storing of the old mfn in page->index to gnttab_map_refs
- move the function header update to a separate patch
v3:
- a new approach to retain old behaviour where it needed
- squash the patches into one
v4:
- move out the common bits from m2p* functions, and pass pfn/mfn as parameter
- clear page->private before doing anything with the page, so m2p_find_override
won't race with this
v5:
- change return value handling in __gnttab_[un]map_refs
- remove a stray space in page.h
- add detail why ret = 0 now at some places
v6:
- don't pass pfn to m2p* functions, just get it locally
Signed-off-by: Zoltan Kiss <zoltan.kiss@citrix.com>
Suggested-by: David Vrabel <david.vrabel@citrix.com>
Acked-by: David Vrabel <david.vrabel@citrix.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Merge misc fixes from Andrew Morton:
"A few hotfixes and various leftovers which were awaiting other merges.
Mainly movement of zram into mm/"
* emailed patches fron Andrew Morton <akpm@linux-foundation.org>: (25 commits)
memcg: fix mutex not unlocked on memcg_create_kmem_cache fail path
Documentation/filesystems/vfs.txt: update file_operations documentation
mm, oom: base root bonus on current usage
mm: don't lose the SOFT_DIRTY flag on mprotect
mm/slub.c: fix page->_count corruption (again)
mm/mempolicy.c: fix mempolicy printing in numa_maps
zram: remove zram->lock in read path and change it with mutex
zram: remove workqueue for freeing removed pending slot
zram: introduce zram->tb_lock
zram: use atomic operation for stat
zram: remove unnecessary free
zram: delay pending free request in read path
zram: fix race between reset and flushing pending work
zsmalloc: add maintainers
zram: add zram maintainers
zsmalloc: add copyright
zram: add copyright
zram: remove old private project comment
zram: promote zram from staging
zsmalloc: move it under mm
...
Pull x86 asmlinkage (LTO) changes from Peter Anvin:
"This patchset adds more infrastructure for link time optimization
(LTO).
This patchset was pulled into my tree late because of a
miscommunication (part of the patchset was picked up by other
maintainers). However, the patchset is strictly build-related and
seems to be okay in testing"
* 'x86-asmlinkage-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, asmlinkage, xen: Fix type of NMI
x86, asmlinkage, xen, kvm: Make {xen,kvm}_lock_spinning global and visible
x86: Use inline assembler instead of global register variable to get sp
x86, asmlinkage, paravirt: Make paravirt thunks global
x86, asmlinkage, paravirt: Don't rely on local assembler labels
x86, asmlinkage, lguest: Fix C functions used by inline assembler
Pull x86 build bits from Peter Anvin:
"Various build-related minor bits.
Most of this is work by David Woodhouse to be able to compile the
early boot code with clang/llvm; we have also managed to push an
actual -m16 option into gcc 4.9 so this makes us use that option if
available instead of hacking it.
The balance is a patch from Michael Davidson to the relocs program to
help manual debugging.
None of these should change the actual compiled binary with currently
released compilers"
* 'x86-build-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, build: Build 16-bit code with -m16 where possible
x86, boot: Fix word-size assumptions in has_eflag() inline asm
x86, boot: Use __attribute__((used)) to ensure videocard structs are emitted
x86: Remove duplication of 16-bit CFLAGS
x86, relocs: Add manual debug mode
Pull __TIME__/__DATE__ removal from Michal Marek:
"This series by Josh finishes the removal of __DATE__ and __TIME__ from
the kernel. The last patch adds -Werror=date-time to KBUILD_CFLAGS to
stop these from reappearing.
Part of the series went through Greg's trees during this merge window,
which is why this pull request is not based on v3.13-rc1"
* 'drop-time' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
Makefile: Build with -Werror=date-time if the compiler supports it
x86: math-emu: Drop already-disabled print of build date
net: wireless: brcm80211: Drop debug version with build date/time
mtd: denali: Drop print of build date/time
The SOFT_DIRTY bit shows that the content of memory was changed after a
defined point in the past. mprotect() doesn't change the content of
memory, so it must not change the SOFT_DIRTY bit.
This bug causes a malfunction: on the first iteration all pages are
dumped. On other iterations only pages with the SOFT_DIRTY bit are
dumped. So if the SOFT_DIRTY bit is cleared from a page by mistake, the
page is not dumped and its content will be restored incorrectly.
This patch does nothing with _PAGE_SWP_SOFT_DIRTY, becase pte_modify()
is called only for present pages.
Fixes commit 0f8975ec4d ("mm: soft-dirty bits for user memory changes
tracking").
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Acked-by: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Wen Congyang <wency@cn.fujitsu.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Further discussion here: http://marc.info/?l=linux-kernel&m=139073901101034&w=2
kbuild, 0day kernel build service, outputs the warning:
arch/x86/kernel/irq.c:333:1: warning: the frame size of 2056 bytes
is larger than 2048 bytes [-Wframe-larger-than=]
because check_irq_vectors_for_cpu_disable() allocates two cpumasks on the
stack. Fix this by moving the two cpumasks to a global file context.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Tested-by: David Rientjes <rientjes@google.com>
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Link: http://lkml.kernel.org/r/1390915331-27375-1-git-send-email-prarit@redhat.com
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Michel Lespinasse <walken@google.com>
Cc: Seiji Aguchi <seiji.aguchi@hds.com>
Cc: Yang Zhang <yang.z.zhang@Intel.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Janet Morgan <janet.morgan@intel.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Ruiv Wang <ruiv.wang@gmail.com>
Cc: Gong Chen <gong.chen@linux.intel.com>
Cc: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Both clang 3.5 and GCC 4.9 will support this (as of r199754 and r207196
respectively). Both have been tested to produce booting kernels when the
16-bit code is built with -m16. (Modulo LLVM PR3997, at least.)
[ hpa: folded test for -m16 into M16_CFLAGS ]
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Link: http://lkml.kernel.org/r/1390997807.20153.133.camel@i7.infradead.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Commit dd78b97367 ("x86, boot: Move CPU
flags out of cpucheck") introduced ambiguous inline asm in the
has_eflag() function. In 16-bit mode want the instruction to be
'pushfl', but we just say 'pushf' and hope the compiler does what we
wanted.
When building with 'clang -m16', it won't, because clang doesn't use
the horrid '.code16gcc' hack that even 'gcc -m16' uses internally.
Say what we mean and don't make the compiler make assumptions.
[ hpa: ideally we would be able to use the gcc %zN construct here, but
that is broken for 64-bit integers in gcc < 4.5.
The code with plain "pushf/popf" is fine for 32- or 64-bit mode, but
not for 16-bit mode; in 16-bit mode those are 16-bit instructions in
.code16 mode, and 32-bit instructions in .code16gcc mode. ]
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Link: http://lkml.kernel.org/r/1391079628.26079.82.camel@shinybook.infradead.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
LTO requires consistent types of symbols over all files.
So "nmi" cannot be declared as a char [] here, need to use the
correct function type.
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/1382458079-24450-8-git-send-email-andi@firstfloor.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
These functions are called from inline assembler stubs, thus
need to be global and visible.
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Gleb Natapov <gleb@kernel.org>
Cc: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/1382458079-24450-7-git-send-email-andi@firstfloor.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
LTO in gcc 4.6/47. has trouble with global register variables. They were used
to read the stack pointer. Use a simple inline assembler statement with
a mov instead.
This also helps LLVM/clang, which does not support global register
variables.
[ hpa: Ideally this should become a builtin in both gcc and clang. ]
v2: More general asm constraint. Fix description (Jan Beulich)
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/1382458079-24450-6-git-send-email-andi@firstfloor.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
The paravirt thunks use a hack of using a static reference to a static
function to reference that function from the top level statement.
This assumes that gcc always generates static function names in a specific
format, which is not necessarily true.
Simply make these functions global and asmlinkage or __visible. This way the
static __used variables are not needed and everything works.
Functions with arguments are __visible to keep the register calling
convention on 32bit.
Changed in paravirt and in all users (Xen and vsmp)
v2: Use __visible for functions with arguments
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Ido Yariv <ido@wizery.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/1382458079-24450-5-git-send-email-andi@firstfloor.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
The paravirt patching code assumes that it can reference a
local assembler label between two different top level assembler
statements. This does not work with LTO
where the assembler code may end up in different assembler files.
Replace it with extern / global /asm linkage labels.
This also removes one redundant copy of the macro.
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/1382458079-24450-4-git-send-email-andi@firstfloor.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
- Make the C code used by the paravirt stubs visible
- Since they have to be global now, give them a more unique
name.
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Link: http://lkml.kernel.org/r/1382458079-24450-3-git-send-email-andi@firstfloor.org
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Pull intel MID cleanups from Peter Anvin:
"Miscellaneous cleanups to the intel-mid code merged earlier in this
merge window"
* 'x86-intel-mid-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, intel-mid: Cleanup some platform code's header files
x86, intel-mid: Add missing 'void' to functions without arguments
x86: Don't add new __cpuinit users to Merrifield platform code
x86: Don't introduce more __cpuinit users in intel_mid_weak_decls.h
Pull more x32 uabi type fixes from Peter Anvin:
"Despite the branch name, **most of these changes are to generic
code**. They change types so that they make an increasing amount of
the exported uapi kernel headers usable for libc.
The ARM64 people are also interested in these changes for their ILP32
ABI"
* 'x86-x32-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
uapi: Use __kernel_long_t in struct mq_attr
uapi: Use __kernel_ulong_t in shmid64_ds/shminfo64/shm_info
x86, uapi, x32: Use __kernel_ulong_t in x86 struct semid64_ds
uapi: Use __kernel_ulong_t in struct msqid64_ds
uapi: Use __kernel_long_t in struct msgbuf
uapi, asm-generic: Use __kernel_ulong_t in uapi struct ipc64_perm
uapi: Use __kernel_long_t/__kernel_ulong_t in <linux/resource.h>
uapi: Use __kernel_long_t in struct timex
When Hyper-V hypervisor leaves are present, KVM must relocate
its own leaves at 0x40000100, because Windows does not look for
Hyper-V leaves at indices other than 0x40000000. In this case,
the KVM features are at 0x40000101, but the old code would always
look at 0x40000001.
Fix by using kvm_cpuid_base(). This also requires making the
function non-inline, since kvm_cpuid_base() is static.
Fixes: 1085ba7f55
Cc: stable@vger.kernel.org
Cc: mtosatti@redhat.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It is unnecessary to go through hypervisor_cpuid_base every time
a leaf is found (which will be every time a feature is requested
after the next patch).
Fixes: 1085ba7f55
Cc: stable@vger.kernel.org
Cc: mtosatti@redhat.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
platform_ipc.h and platform_msic.h are wrongly declaring functions as
external and with 'weak' attribute. This patch does a cleanup on those
header files.
It should have no functional change.
Signed-off-by: David Cohen <david.a.cohen@linux.intel.com>
Link: http://lkml.kernel.org/r/1390950567-12821-1-git-send-email-david.a.cohen@linux.intel.com
Cc: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Dave reported big numa system booting is broken.
It turns out that commit 5b6e529521 ("x86: memblock: set current limit
to max low memory address") sets the limit to low wrongly.
max_low_pfn_mapped is different from max_pfn_mapped.
max_low_pfn_mapped is always under 4G.
That will memblock_alloc_nid all go under 4G.
Revert 5b6e529521 to fix a no-boot regression which was triggered by
457ff1de2d ("lib/swiotlb.c: use memblock apis for early memory
allocations").
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Reported-by: Dave Hansen <dave.hansen@intel.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The kernel already has this information, so other bits of kernel code
shouldn't duplicate that. This also eliminates the use of __DATE__,
which makes the build non-deterministic.
This message was already disabled at build time, with PRINT_MESSAGES
undef'd at the top of the file.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Michal Marek <mmarek@suse.cz>
Check for invalid state transitions on guest-initiated updates of
MSR_IA32_APICBASE. This address both enabling of the x2APIC when it is
not supported and all invalid transitions as described in SDM section
10.12.5. It also checks that no reserved bit is set in APICBASE by the
guest.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
[Use cpuid_maxphyaddr instead of guest_cpuid_get_phys_bits. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Pull networking updates from David Miller:
1) BPF debugger and asm tool by Daniel Borkmann.
2) Speed up create/bind in AF_PACKET, also from Daniel Borkmann.
3) Correct reciprocal_divide and update users, from Hannes Frederic
Sowa and Daniel Borkmann.
4) Currently we only have a "set" operation for the hw timestamp socket
ioctl, add a "get" operation to match. From Ben Hutchings.
5) Add better trace events for debugging driver datapath problems, also
from Ben Hutchings.
6) Implement auto corking in TCP, from Eric Dumazet. Basically, if we
have a small send and a previous packet is already in the qdisc or
device queue, defer until TX completion or we get more data.
7) Allow userspace to manage ipv6 temporary addresses, from Jiri Pirko.
8) Add a qdisc bypass option for AF_PACKET sockets, from Daniel
Borkmann.
9) Share IP header compression code between Bluetooth and IEEE802154
layers, from Jukka Rissanen.
10) Fix ipv6 router reachability probing, from Jiri Benc.
11) Allow packets to be captured on macvtap devices, from Vlad Yasevich.
12) Support tunneling in GRO layer, from Jerry Chu.
13) Allow bonding to be configured fully using netlink, from Scott
Feldman.
14) Allow AF_PACKET users to obtain the VLAN TPID, just like they can
already get the TCI. From Atzm Watanabe.
15) New "Heavy Hitter" qdisc, from Terry Lam.
16) Significantly improve the IPSEC support in pktgen, from Fan Du.
17) Allow ipv4 tunnels to cache routes, just like sockets. From Tom
Herbert.
18) Add Proportional Integral Enhanced packet scheduler, from Vijay
Subramanian.
19) Allow openvswitch to mmap'd netlink, from Thomas Graf.
20) Key TCP metrics blobs also by source address, not just destination
address. From Christoph Paasch.
21) Support 10G in generic phylib. From Andy Fleming.
22) Try to short-circuit GRO flow compares using device provided RX
hash, if provided. From Tom Herbert.
The wireless and netfilter folks have been busy little bees too.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (2064 commits)
net/cxgb4: Fix referencing freed adapter
ipv6: reallocate addrconf router for ipv6 address when lo device up
fib_frontend: fix possible NULL pointer dereference
rtnetlink: remove IFLA_BOND_SLAVE definition
rtnetlink: remove check for fill_slave_info in rtnl_have_link_slave_info
qlcnic: update version to 5.3.55
qlcnic: Enhance logic to calculate msix vectors.
qlcnic: Refactor interrupt coalescing code for all adapters.
qlcnic: Update poll controller code path
qlcnic: Interrupt code cleanup
qlcnic: Enhance Tx timeout debugging.
qlcnic: Use bool for rx_mac_learn.
bonding: fix u64 division
rtnetlink: add missing IFLA_BOND_AD_INFO_UNSPEC
sfc: Use the correct maximum TX DMA ring size for SFC9100
Add Shradha Shah as the sfc driver maintainer.
net/vxlan: Share RX skb de-marking and checksum checks with ovs
tulip: cleanup by using ARRAY_SIZE()
ip_tunnel: clear IPCB in ip_tunnel_xmit() in case dst_link_failure() is called
net/cxgb4: Don't retrieve stats during recovery
...
Pull scheduler fixes from Ingo Molnar:
"A couple of regression fixes mostly hitting virtualized setups, but
also some bare metal systems"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/x86/tsc: Initialize multiplier to 0
sched/clock: Fixup early initialization
sched/preempt/x86: Fix voluntary preempt for x86
Revert "sched: Fix sleep time double accounting in enqueue entity"
This patch fixes the following warning:
warning: (X86_INTEL_MID) selects INTEL_SCU_IPC which has unmet direct dependencies (X86 && X86_PLATFORM_DEVICES && X86_INTEL_MID)
It happens because when selected, X86_INTEL_MID tries to select
INTEL_SCU_IPC regardless all its dependencies are met or not.
This patch fixes it by adding the missing X86_PLATFORM_DEVICES
dependency to X86_INTEL_MID.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: David Cohen <david.a.cohen@linux.intel.com>
Link: http://lkml.kernel.org/r/1390329699-20782-1-git-send-email-david.a.cohen@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
When ACPI SLIT table has an I/O locality (i.e. a locality
unique to an I/O device), numa_set_distance() emits this warning
message:
NUMA: Warning: node ids are out of bound, from=-1 to=-1 distance=10
acpi_numa_slit_init() calls numa_set_distance() with
pxm_to_node(), which assumes that all localities have been
parsed with SRAT previously. SRAT does not list I/O localities,
where as SLIT lists all localities including I/Os. Hence,
pxm_to_node() returns NUMA_NO_NODE (-1) for an I/O locality.
I/O localities are not supported and are ignored today, but emitting
such warning message leads to unnecessary confusion.
Change acpi_numa_slit_init() to avoid calling
numa_set_distance() with NUMA_NO_NODE.
Signed-off-by: Toshi Kani <toshi.kani@hp.com>
Acked-by: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Link: http://lkml.kernel.org/n/tip-dSvpjjvp8aMzs1ybkftxohlh@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There was a large ebizzy performance regression that was
bisected to commit 611ae8e3 (x86/tlb: enable tlb flush range
support for x86). The problem was related to the
tlb_flushall_shift tuning for IvyBridge which was altered. The
problem is that it is not clear if the tuning values for each
CPU family is correct as the methodology used to tune the values
is unclear.
This patch uses a conservative tlb_flushall_shift value for all
CPU families except IvyBridge so the decision can be revisited
if any regression is found as a result of this change.
IvyBridge is an exception as testing with one methodology
determined that the value of 2 is acceptable. Details are in
the changelog for the patch "x86: mm: Change tlb_flushall_shift
for IvyBridge".
One important aspect of this to watch out for is Xen. The
original commit log mentioned large performance gains on Xen.
It's possible Xen is more sensitive to this value if it flushes
small ranges of pages more frequently than workloads on bare
metal typically do.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Tested-by: Davidlohr Bueso <davidlohr@hp.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alex Shi <alex.shi@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-dyzMww3fqugnhbhgo6Gxmtkw@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There was a large performance regression that was bisected to
commit 611ae8e3 ("x86/tlb: enable tlb flush range support for
x86"). This patch simply changes the default balance point
between a local and global flush for IvyBridge.
In the interest of allowing the tests to be reproduced, this
patch was tested using mmtests 0.15 with the following
configurations
configs/config-global-dhp__tlbflush-performance
configs/config-global-dhp__scheduler-performance
configs/config-global-dhp__network-performance
Results are from two machines
Ivybridge 4 threads: Intel(R) Core(TM) i3-3240 CPU @ 3.40GHz
Ivybridge 8 threads: Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz
Page fault microbenchmark showed nothing interesting.
Ebizzy was configured to run multiple iterations and threads.
Thread counts ranged from 1 to NR_CPUS*2. For each thread count,
it ran 100 iterations and each iteration lasted 10 seconds.
Ivybridge 4 threads
3.13.0-rc7 3.13.0-rc7
vanilla altshift-v3
Mean 1 6395.44 ( 0.00%) 6789.09 ( 6.16%)
Mean 2 7012.85 ( 0.00%) 8052.16 ( 14.82%)
Mean 3 6403.04 ( 0.00%) 6973.74 ( 8.91%)
Mean 4 6135.32 ( 0.00%) 6582.33 ( 7.29%)
Mean 5 6095.69 ( 0.00%) 6526.68 ( 7.07%)
Mean 6 6114.33 ( 0.00%) 6416.64 ( 4.94%)
Mean 7 6085.10 ( 0.00%) 6448.51 ( 5.97%)
Mean 8 6120.62 ( 0.00%) 6462.97 ( 5.59%)
Ivybridge 8 threads
3.13.0-rc7 3.13.0-rc7
vanilla altshift-v3
Mean 1 7336.65 ( 0.00%) 7787.02 ( 6.14%)
Mean 2 8218.41 ( 0.00%) 9484.13 ( 15.40%)
Mean 3 7973.62 ( 0.00%) 8922.01 ( 11.89%)
Mean 4 7798.33 ( 0.00%) 8567.03 ( 9.86%)
Mean 5 7158.72 ( 0.00%) 8214.23 ( 14.74%)
Mean 6 6852.27 ( 0.00%) 7952.45 ( 16.06%)
Mean 7 6774.65 ( 0.00%) 7536.35 ( 11.24%)
Mean 8 6510.50 ( 0.00%) 6894.05 ( 5.89%)
Mean 12 6182.90 ( 0.00%) 6661.29 ( 7.74%)
Mean 16 6100.09 ( 0.00%) 6608.69 ( 8.34%)
Ebizzy hits the worst case scenario for TLB range flushing every
time and it shows for these Ivybridge CPUs at least that the
default choice is a poor on. The patch addresses the problem.
Next was a tlbflush microbenchmark written by Alex Shi at
http://marc.info/?l=linux-kernel&m=133727348217113 . It
measures access costs while the TLB is being flushed. The
expectation is that if there are always full TLB flushes that
the benchmark would suffer and it benefits from range flushing
There are 320 iterations of the test per thread count. The
number of entries is randomly selected with a min of 1 and max
of 512. To ensure a reasonably even spread of entries, the full
range is broken up into 8 sections and a random number selected
within that section.
iteration 1, random number between 0-64
iteration 2, random number between 64-128 etc
This is still a very weak methodology. When you do not know
what are typical ranges, random is a reasonable choice but it
can be easily argued that the opimisation was for smaller ranges
and an even spread is not representative of any workload that
matters. To improve this, we'd need to know the probability
distribution of TLB flush range sizes for a set of workloads
that are considered "common", build a synthetic trace and feed
that into this benchmark. Even that is not perfect because it
would not account for the time between flushes but there are
limits of what can be reasonably done and still be doing
something useful. If a representative synthetic trace is
provided then this benchmark could be revisited and the shift values retuned.
Ivybridge 4 threads
3.13.0-rc7 3.13.0-rc7
vanilla altshift-v3
Mean 1 10.50 ( 0.00%) 10.50 ( 0.03%)
Mean 2 17.59 ( 0.00%) 17.18 ( 2.34%)
Mean 3 22.98 ( 0.00%) 21.74 ( 5.41%)
Mean 5 47.13 ( 0.00%) 46.23 ( 1.92%)
Mean 8 43.30 ( 0.00%) 42.56 ( 1.72%)
Ivybridge 8 threads
3.13.0-rc7 3.13.0-rc7
vanilla altshift-v3
Mean 1 9.45 ( 0.00%) 9.36 ( 0.93%)
Mean 2 9.37 ( 0.00%) 9.70 ( -3.54%)
Mean 3 9.36 ( 0.00%) 9.29 ( 0.70%)
Mean 5 14.49 ( 0.00%) 15.04 ( -3.75%)
Mean 8 41.08 ( 0.00%) 38.73 ( 5.71%)
Mean 13 32.04 ( 0.00%) 31.24 ( 2.49%)
Mean 16 40.05 ( 0.00%) 39.04 ( 2.51%)
For both CPUs, average access time is reduced which is good as
this is the benchmark that was used to tune the shift values in
the first place albeit it is now known *how* the benchmark was
used.
The scheduler benchmarks were somewhat inconclusive. They
showed gains and losses and makes me reconsider how stable those
benchmarks really are or if something else might be interfering
with the test results recently.
Network benchmarks were inconclusive. Almost all results were
flat except for netperf-udp tests on the 4 thread machine.
These results were unstable and showed large variations between
reboots. It is unknown if this is a recent problems but I've
noticed before that netperf-udp results tend to vary.
Based on these results, changing the default for Ivybridge seems
like a logical choice.
Signed-off-by: Mel Gorman <mgorman@suse.de>
Tested-by: Davidlohr Bueso <davidlohr@hp.com>
Reviewed-by: Alex Shi <alex.shi@linaro.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-cqnadffh1tiqrshthRj3Esge@git.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
