Rewrite the vmap allocator to use rbtrees and lazy tlb flushing, and
provide a fast, scalable percpu frontend for small vmaps (requires a
slightly different API, though).
The biggest problem with vmap is actually vunmap. Presently this requires
a global kernel TLB flush, which on most architectures is a broadcast IPI
to all CPUs to flush the cache. This is all done under a global lock. As
the number of CPUs increases, so will the number of vunmaps a scaled
workload will want to perform, and so will the cost of a global TLB flush.
This gives terrible quadratic scalability characteristics.
Another problem is that the entire vmap subsystem works under a single
lock. It is a rwlock, but it is actually taken for write in all the fast
paths, and the read locking would likely never be run concurrently anyway,
so it's just pointless.
This is a rewrite of vmap subsystem to solve those problems. The existing
vmalloc API is implemented on top of the rewritten subsystem.
The TLB flushing problem is solved by using lazy TLB unmapping. vmap
addresses do not have to be flushed immediately when they are vunmapped,
because the kernel will not reuse them again (would be a use-after-free)
until they are reallocated. So the addresses aren't allocated again until
a subsequent TLB flush. A single TLB flush then can flush multiple
vunmaps from each CPU.
XEN and PAT and such do not like deferred TLB flushing because they can't
always handle multiple aliasing virtual addresses to a physical address.
They now call vm_unmap_aliases() in order to flush any deferred mappings.
That call is very expensive (well, actually not a lot more expensive than
a single vunmap under the old scheme), however it should be OK if not
called too often.
The virtual memory extent information is stored in an rbtree rather than a
linked list to improve the algorithmic scalability.
There is a per-CPU allocator for small vmaps, which amortizes or avoids
global locking.
To use the per-CPU interface, the vm_map_ram / vm_unmap_ram interfaces
must be used in place of vmap and vunmap. Vmalloc does not use these
interfaces at the moment, so it will not be quite so scalable (although it
will use lazy TLB flushing).
As a quick test of performance, I ran a test that loops in the kernel,
linearly mapping then touching then unmapping 4 pages. Different numbers
of tests were run in parallel on an 4 core, 2 socket opteron. Results are
in nanoseconds per map+touch+unmap.
threads vanilla vmap rewrite
1 14700 2900
2 33600 3000
4 49500 2800
8 70631 2900
So with a 8 cores, the rewritten version is already 25x faster.
In a slightly more realistic test (although with an older and less
scalable version of the patch), I ripped the not-very-good vunmap batching
code out of XFS, and implemented the large buffer mapping with vm_map_ram
and vm_unmap_ram... along with a couple of other tricks, I was able to
speed up a large directory workload by 20x on a 64 CPU system. I believe
vmap/vunmap is actually sped up a lot more than 20x on such a system, but
I'm running into other locks now. vmap is pretty well blown off the
profiles.
Before:
1352059 total 0.1401
798784 _write_lock 8320.6667 <- vmlist_lock
529313 default_idle 1181.5022
15242 smp_call_function 15.8771 <- vmap tlb flushing
2472 __get_vm_area_node 1.9312 <- vmap
1762 remove_vm_area 4.5885 <- vunmap
316 map_vm_area 0.2297 <- vmap
312 kfree 0.1950
300 _spin_lock 3.1250
252 sn_send_IPI_phys 0.4375 <- tlb flushing
238 vmap 0.8264 <- vmap
216 find_lock_page 0.5192
196 find_next_bit 0.3603
136 sn2_send_IPI 0.2024
130 pio_phys_write_mmr 2.0312
118 unmap_kernel_range 0.1229
After:
78406 total 0.0081
40053 default_idle 89.4040
33576 ia64_spinlock_contention 349.7500
1650 _spin_lock 17.1875
319 __reg_op 0.5538
281 _atomic_dec_and_lock 1.0977
153 mutex_unlock 1.5938
123 iget_locked 0.1671
117 xfs_dir_lookup 0.1662
117 dput 0.1406
114 xfs_iget_core 0.0268
92 xfs_da_hashname 0.1917
75 d_alloc 0.0670
68 vmap_page_range 0.0462 <- vmap
58 kmem_cache_alloc 0.0604
57 memset 0.0540
52 rb_next 0.1625
50 __copy_user 0.0208
49 bitmap_find_free_region 0.2188 <- vmap
46 ia64_sn_udelay 0.1106
45 find_inode_fast 0.1406
42 memcmp 0.2188
42 finish_task_switch 0.1094
42 __d_lookup 0.0410
40 radix_tree_lookup_slot 0.1250
37 _spin_unlock_irqrestore 0.3854
36 xfs_bmapi 0.0050
36 kmem_cache_free 0.0256
35 xfs_vn_getattr 0.0322
34 radix_tree_lookup 0.1062
33 __link_path_walk 0.0035
31 xfs_da_do_buf 0.0091
30 _xfs_buf_find 0.0204
28 find_get_page 0.0875
27 xfs_iread 0.0241
27 __strncpy_from_user 0.2812
26 _xfs_buf_initialize 0.0406
24 _xfs_buf_lookup_pages 0.0179
24 vunmap_page_range 0.0250 <- vunmap
23 find_lock_page 0.0799
22 vm_map_ram 0.0087 <- vmap
20 kfree 0.0125
19 put_page 0.0330
18 __kmalloc 0.0176
17 xfs_da_node_lookup_int 0.0086
17 _read_lock 0.0885
17 page_waitqueue 0.0664
vmap has gone from being the top 5 on the profiles and flushing the crap
out of all TLBs, to using less than 1% of kernel time.
[akpm@linux-foundation.org: cleanups, section fix]
[akpm@linux-foundation.org: fix build on alpha]
Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Krzysztof Helt <krzysztof.h1@poczta.fm>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
There is nothing architecture specific about remove_memory().
remove_memory() function is common for all architectures which support
hotplug memory remove. Instead of duplicating it in every architecture,
collapse them into arch neutral function.
[akpm@linux-foundation.org: fix the export]
Signed-off-by: Badari Pulavarty <pbadari@us.ibm.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Gary Hade <garyhade@us.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The driver would like to map IO space directly for copying data in when
appropriate, to avoid CPU cache flushing for streaming writes.
kmap_atomic_pfn lets us avoid IPIs associated with ioremap for this process.
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
* 'kvm-updates/2.6.28' of git://git.kernel.org/pub/scm/linux/kernel/git/avi/kvm: (134 commits)
KVM: ia64: Add intel iommu support for guests.
KVM: ia64: add directed mmio range support for kvm guests
KVM: ia64: Make pmt table be able to hold physical mmio entries.
KVM: Move irqchip_in_kernel() from ioapic.h to irq.h
KVM: Separate irq ack notification out of arch/x86/kvm/irq.c
KVM: Change is_mmio_pfn to kvm_is_mmio_pfn, and make it common for all archs
KVM: Move device assignment logic to common code
KVM: Device Assignment: Move vtd.c from arch/x86/kvm/ to virt/kvm/
KVM: VMX: enable invlpg exiting if EPT is disabled
KVM: x86: Silence various LAPIC-related host kernel messages
KVM: Device Assignment: Map mmio pages into VT-d page table
KVM: PIC: enhance IPI avoidance
KVM: MMU: add "oos_shadow" parameter to disable oos
KVM: MMU: speed up mmu_unsync_walk
KVM: MMU: out of sync shadow core
KVM: MMU: mmu_convert_notrap helper
KVM: MMU: awareness of new kvm_mmu_zap_page behaviour
KVM: MMU: mmu_parent_walk
KVM: x86: trap invlpg
KVM: MMU: sync roots on mmu reload
...
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86: fix compat-vdso
x86/mm: unify init task OOM handling
x86/mm: do not trigger a kernel warning if user-space disables interrupts and generates a page fault
As noted by Akinobu Mita in patch b1fceac2b9,
alloc_bootmem and related functions never return NULL and always return a
zeroed region of memory. Thus a NULL test or memset after calls to these
functions is unnecessary.
arch/frv/mm/init.c | 2 --
1 file changed, 2 deletions(-)
This was fixed using the following semantic patch.
(http://www.emn.fr/x-info/coccinelle/)
// <smpl>
@@
expression E;
statement S;
@@
E = \(alloc_bootmem\|alloc_bootmem_low\|alloc_bootmem_pages\|alloc_bootmem_low_pages\)(...)
... when != E
(
- BUG_ON (E == NULL);
|
- if (E == NULL) S
)
@@
expression E,E1;
@@
E = \(alloc_bootmem\|alloc_bootmem_low\|alloc_bootmem_pages\|alloc_bootmem_low_pages\)(...)
... when != E
- memset(E,0,E1);
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Provide dma_map_page() for the NOMMU-mode FRV arch.
Also do some fixing on the comments attached to the various DMA functions for
both MMU and NOMMU mode FRV code.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Use the generic pci_enable_resources() instead of the arch-specific code.
Unlike this arch-specific code, the generic version:
- checks PCI_NUM_RESOURCES (11), not 6, resources
- skips resources that have neither IORESOURCE_IO nor IORESOURCE_MEM set
- skips ROM resources unless IORESOURCE_ROM_ENABLE is set
- checks for resource collisions with "!r->parent"
Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cooloney/blackfin-2.6:
Blackfin arch: make sure cycles is marked as volatile so gcc doesnt reorder on us
Blackfin arch: disable CONFIG_HW_RANDOM and CONFIG_DAB in defconfig files
Blackfin arch: update cache flush prototypes with argument names to make them less mysterious
Blackfin arch: move bfin_addr_dcachable() and friends into the cacheflush header where it belongs
Blackfin arch: use the new bfin_addr_dcachable() function
Blackfin arch: fix bug - build kernel failed at head.S when reprogram clock on all platforms
Blackfin arch: unify/cleanup cache code
Blackfin arch: update AD7879 platform resources in board file
Blackfin arch: Zero out bss region in L1/L2 memory.
Blackfin arch: add read/write IO accessor functions to Blackfin
Blackfin arch: fix bug - some serial header files set RTS to an input when they should all be outputs
* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6: (46 commits)
UIO: Fix mapping of logical and virtual memory
UIO: add automata sercos3 pci card support
UIO: Change driver name of uio_pdrv
UIO: Add alignment warnings for uio-mem
Driver core: add bus_sort_breadthfirst() function
NET: convert the phy_device file to use bus_find_device_by_name
kobject: Cleanup kobject_rename and !CONFIG_SYSFS
kobject: Fix kobject_rename and !CONFIG_SYSFS
sysfs: Make dir and name args to sysfs_notify() const
platform: add new device registration helper
sysfs: use ilookup5() instead of ilookup5_nowait()
PNP: create device attributes via default device attributes
Driver core: make bus_find_device_by_name() more robust
usb: turn dev_warn+WARN_ON combos into dev_WARN
debug: use dev_WARN() rather than WARN_ON() in device_pm_add()
debug: Introduce a dev_WARN() function
sysfs: fix deadlock
device model: Do a quickcheck for driver binding before doing an expensive check
Driver core: Fix cleanup in device_create_vargs().
Driver core: Clarify device cleanup.
...
name and nlen parameters passed to ->strategy hook are unused, remove
them. In general ->strategy hook should know what it's doing, and don't
do something tricky for which, say, pointer to original userspace array
may be needed (name).
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net> [ networking bits ]
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: David Howells <dhowells@redhat.com>
Cc: Matt Mackall <mpm@selenic.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently, it is possible to set a graphics VESA mode at boot time via the
vga= parameter even when no framebuffer driver supporting this is
configured. This could lead to the system booting with a black screen,
without a usable console.
Fix this problem by only allowing to set graphics modes at boot time if a
supporting framebuffer driver is configured.
Signed-off-by: Michal Januszewski <spock@gentoo.org>
Acked-by: Krzysztof Helt <krzysztof.h1@wp.pl>
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>
Add a pin configuration callback for the s3c24xx SPI driver, as there are
several options depending on the channel and the chip in use.
This is needed as the controller may not have been setup by the initial
bootloader and the fact that the SPI controller gets reset over
suspend/resume into slave mode but the GPIO function registers do not.
Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This change is Alpha-specific. It adds field 'taso' into struct
linux_binprm to remember if the application is TASO. Previously, field
sh_bang was used for this purpose.
Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Pavel Emelyanov <xemul@openvz.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is a preparation patch for introducing a generic iommu_num_pages function.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is a preparation patch for introducing a generic iommu_num_pages function.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Acked-by: David S. Miller <davem@davemloft.net>
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This series of patches re-introduces the iommu_num_pages function so that
it can be used by each architecture specific IOMMU implementations. The
series also changes IOMMU implementations for X86, Alpha, PowerPC and
UltraSparc. The other implementations are not yet changed because the
modifications required are not obvious and I can't test them on real
hardware.
This patch:
This is a preparation patch for introducing a generic iommu_num_pages function.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: Muli Ben-Yehuda <muli@il.ibm.com>
Cc: Dave Airlie <airlied@linux.ie>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Nothing arch specific in get/settimeofday. The details of the timeval
conversion varied a little from arch to arch, but all with the same
results.
Also add an extern declaration for sys_tz to linux/time.h because externs
in .c files are fowned upon. I'll kill the externs in various other files
in a sparate patch.
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: David S. Miller <davem@davemloft.net> [ sparc bits ]
Cc: "Luck, Tony" <tony.luck@intel.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Acked-by: Kyle McMartin <kyle@mcmartin.ca>
Cc: Matthew Wilcox <matthew@wil.cx>
Cc: Grant Grundler <grundler@parisc-linux.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.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>
struct stat / compat_stat is the same on all architectures, so
cp_compat_stat should be, too.
Turns out it is, except that various architectures have slightly and some
high2lowuid/high2lowgid or the direct assignment instead of the
SET_UID/SET_GID that expands to the correct one anyway.
This patch replaces the arch-specific cp_compat_stat implementations with
a common one based on the x86-64 one.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: David S. Miller <davem@davemloft.net> [ sparc bits ]
Acked-by: Kyle McMartin <kyle@mcmartin.ca> [ parisc bits ]
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It's somewhat unlikely that it happens, but right now a race window
between interrupts or machine checks or oopses could corrupt the tainted
bitmap because it is modified in a non atomic fashion.
Convert the taint variable to an unsigned long and use only atomic bit
operations on it.
Unfortunately this means the intvec sysctl functions cannot be used on it
anymore.
It turned out the taint sysctl handler could actually be simplified a bit
(since it only increases capabilities) so this patch actually removes
code.
[akpm@linux-foundation.org: remove unneeded include]
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Using "def_bool n" is pointless, simply using bool here appears more
appropriate.
Further, retaining such options that don't have a prompt and aren't
selected by anything seems also at least questionable.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
__FUNCTION__ is gcc-specific, use __func__
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Remove the dead CONFIG_TTY_LOG (no kconfig option).
Reported-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Cc: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Remove a dead URL.
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Remove a CVS keyword that wasn't updated for a long time from a comment.
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Print the name of the last-accessed sysfs file when we oops, to help track
down oopses which occur in sysfs store/read handlers. Because these oopses
tend to not leave any trace of the offending code in the stack traces.
Cc: Kay Sievers <kay.sievers@vrfy.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
- to be correct wrt to end ranges
- to be optimal with a one-instruction hardware loop
Signed-off-by: Mike Frysinger <vapier.adi@gmail.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
