Make the following style cleanups:
* drop unnecessary //#include from xen-asm_32.S
* compulsive adding of space after comma
* reformat multiline comments
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Enable the use of the direct vcpu-access operations on 64-bit.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Now that x86-64 has directly accessible percpu variables, it can also
implement the direct versions of these operations, which operate on a
vcpu_info structure directly embedded in the percpu area.
In fact, the 64-bit versions are more or less identical, and so can be
shared. The only two differences are:
1. xen_restore_fl_direct takes its argument in eax on 32-bit, and rdi on 64-bit.
Unfortunately it isn't possible to directly refer to the 2nd lsb of rdi directly
(as you can with %ah), so the code isn't quite as dense.
2. check_events needs to variants to save different registers.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
We need to access percpu data fairly early, so set up the percpu
registers as soon as possible. We only need to load the appropriate
segment register. We already have a GDT, but its hard to change it
early because we need to manipulate the pagetable to do so, and that
hasn't been set up yet.
Also, set the kernel stack when bringing up secondary CPUs. If we
don't they all end up sharing the same stack...
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Moving the mmu code from enlighten.c to mmu.c inadvertently broke the
32-bit build. Fix it.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Impact: Fix race condition
xen_mc_batch has a small preempt race where it takes the address of a
percpu variable immediately before disabling interrupts, thereby
leaving a small window in which we may migrate to another cpu and save
the flags in the wrong percpu variable. Disable interrupts before
saving the old flags in a percpu.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Impact: fix xen booting
We need to access percpu data fairly early, so set up the percpu
registers as soon as possible. We only need to load the appropriate
segment register. We already have a GDT, but its hard to change it
early because we need to manipulate the pagetable to do so, and that
hasn't been set up yet.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Impact: Optimization
In the native case, pte_val, make_pte, etc are all just identity
functions, so there's no need to clobber a lot of registers over them.
(This changes the 32-bit callee-save calling convention to return both
EAX and EDX so functions can return 64-bit values.)
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Impact: Optimization
One of the problems with inserting a pile of C calls where previously
there were none is that the register pressure is greatly increased.
The C calling convention says that the caller must expect a certain
set of registers may be trashed by the callee, and that the callee can
use those registers without restriction. This includes the function
argument registers, and several others.
This patch seeks to alleviate this pressure by introducing wrapper
thunks that will do the register saving/restoring, so that the
callsite doesn't need to worry about it, but the callee function can
be conventional compiler-generated code. In many cases (particularly
performance-sensitive cases) the callee will be in assembler anyway,
and need not use the compiler's calling convention.
Standard calling convention is:
arguments return scratch
x86-32 eax edx ecx eax ?
x86-64 rdi rsi rdx rcx rax r8 r9 r10 r11
The thunk preserves all argument and scratch registers. The return
register is not preserved, and is available as a scratch register for
unwrapped callee code (and of course the return value).
Wrapped function pointers are themselves wrapped in a struct
paravirt_callee_save structure, in order to get some warning from the
compiler when functions with mismatched calling conventions are used.
The most common paravirt ops, both statically and dynamically, are
interrupt enable/disable/save/restore, so handle them first. This is
particularly easy since their calls are handled specially anyway.
XXX Deal with VMI. What's their calling convention?
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Impact: Cleanup
Move remaining mmu-related stuff into mmu.c.
A general cleanup, and lay the groundwork for later patches.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Impact: cleanup
Rename init_gdt() to setup_percpu_segment(), and move it to
setup_percpu.c.
Signed-off-by: Brian Gerst <brgerst@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Impact: build fix
This build error:
arch/x86/xen/suspend.c:22: error: implicit declaration of function 'fix_to_virt'
arch/x86/xen/suspend.c:22: error: 'FIX_PARAVIRT_BOOTMAP' undeclared (first use in this function)
arch/x86/xen/suspend.c:22: error: (Each undeclared identifier is reported only once
arch/x86/xen/suspend.c:22: error: for each function it appears in.)
triggers because the hardirq.h unification removed an implicit fixmap.h
include - on which arch/x86/xen/suspend.c depended. Add the fixmap.h
include explicitly.
Signed-off-by: Ingo Molnar <mingo@elte.hu>
pte_flags() was introduced as a new pvop in order to extract just the
flags portion of a pte, which is a potentially cheaper operation than
extracting the page number as well. It turns out this operation is
not needed, because simply using a mask to extract the flags from a
pte is sufficient for all current users.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: cleanup
Copy the code to cpu_init() to satisfy the requirement that the cpu
be reinitialized. Remove all other calls, since the segments are
already initialized in head_64.S.
Signed-off-by: Brian Gerst <brgerst@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
tj: * in asm-offsets_64.c, pda.h inclusion shouldn't be removed as pda
is still referenced in the file
* s/oldrsp/old_rsp/
Signed-off-by: Brian Gerst <brgerst@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Also clean up PER_CPU_VAR usage in xen-asm_64.S
tj: * remove now unused stack_thread_info()
* s/kernelstack/kernel_stack/
* added FIXME comment in xen-asm_64.S
Signed-off-by: Brian Gerst <brgerst@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
It is an optimization and a cleanup, and adds the following new
generic percpu methods:
percpu_read()
percpu_write()
percpu_add()
percpu_sub()
percpu_and()
percpu_or()
percpu_xor()
and implements support for them on x86. (other architectures will fall
back to a default implementation)
The advantage is that for example to read a local percpu variable,
instead of this sequence:
return __get_cpu_var(var);
ffffffff8102ca2b: 48 8b 14 fd 80 09 74 mov -0x7e8bf680(,%rdi,8),%rdx
ffffffff8102ca32: 81
ffffffff8102ca33: 48 c7 c0 d8 59 00 00 mov $0x59d8,%rax
ffffffff8102ca3a: 48 8b 04 10 mov (%rax,%rdx,1),%rax
We can get a single instruction by using the optimized variants:
return percpu_read(var);
ffffffff8102ca3f: 65 48 8b 05 91 8f fd mov %gs:0x7efd8f91(%rip),%rax
I also cleaned up the x86-specific APIs and made the x86 code use
these new generic percpu primitives.
tj: * fixed generic percpu_sub() definition as Roel Kluin pointed out
* added percpu_and() for completeness's sake
* made generic percpu ops atomic against preemption
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Tejun Heo <tj@kernel.org>
Do the following cleanups:
* kill x86_64_init_pda() which now is equivalent to pda_init()
* use per_cpu_offset() instead of cpu_pda() when initializing
initial_gs
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
[ Based on original patch from Christoph Lameter and Mike Travis. ]
Currently pdas and percpu areas are allocated separately. %gs points
to local pda and percpu area can be reached using pda->data_offset.
This patch folds pda into percpu area.
Due to strange gcc requirement, pda needs to be at the beginning of
the percpu area so that pda->stack_canary is at %gs:40. To achieve
this, a new percpu output section macro - PERCPU_VADDR_PREALLOC() - is
added and used to reserve pda sized chunk at the start of the percpu
area.
After this change, for boot cpu, %gs first points to pda in the
data.init area and later during setup_per_cpu_areas() gets updated to
point to the actual pda. This means that setup_per_cpu_areas() need
to reload %gs for CPU0 while clearing pda area for other cpus as cpu0
already has modified it when control reaches setup_per_cpu_areas().
This patch also removes now unnecessary get_local_pda() and its call
sites.
A lot of this patch is taken from Mike Travis' "x86_64: Fold pda into
per cpu area" patch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: reduce stack usage, use new cpumask API.
This is made a little more tricky by uv_flush_tlb_others which
actually alters its argument, for an IPI to be sent to the remaining
cpus in the mask.
I solve this by allocating a cpumask_var_t for this case and falling back
to IPI should this fail.
To eliminate temporaries in the caller, all flush_tlb_others implementations
now do the this-cpu-elimination step themselves.
Note also the curious "cpus_or(f->flush_cpumask, cpumask, f->flush_cpumask)"
which has been there since pre-git and yet f->flush_cpumask is always zero
at this point.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
* 'cpus4096-for-linus-2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (66 commits)
x86: export vector_used_by_percpu_irq
x86: use logical apicid in x2apic_cluster's x2apic_cpu_mask_to_apicid_and()
sched: nominate preferred wakeup cpu, fix
x86: fix lguest used_vectors breakage, -v2
x86: fix warning in arch/x86/kernel/io_apic.c
sched: fix warning in kernel/sched.c
sched: move test_sd_parent() to an SMP section of sched.h
sched: add SD_BALANCE_NEWIDLE at MC and CPU level for sched_mc>0
sched: activate active load balancing in new idle cpus
sched: bias task wakeups to preferred semi-idle packages
sched: nominate preferred wakeup cpu
sched: favour lower logical cpu number for sched_mc balance
sched: framework for sched_mc/smt_power_savings=N
sched: convert BALANCE_FOR_xx_POWER to inline functions
x86: use possible_cpus=NUM to extend the possible cpus allowed
x86: fix cpu_mask_to_apicid_and to include cpu_online_mask
x86: update io_apic.c to the new cpumask code
x86: Introduce topology_core_cpumask()/topology_thread_cpumask()
x86: xen: use smp_call_function_many()
x86: use work_on_cpu in x86/kernel/cpu/mcheck/mce_amd_64.c
...
Fixed up trivial conflict in kernel/time/tick-sched.c manually
The cpu time spent by the idle process actually doing something is
currently accounted as idle time. This is plain wrong, the architectures
that support VIRT_CPU_ACCOUNTING=y can do better: distinguish between the
time spent doing nothing and the time spent by idle doing work. The first
is accounted with account_idle_time and the second with account_system_time.
The architectures that use the account_xxx_time interface directly and not
the account_xxx_ticks interface now need to do the check for the idle
process in their arch code. In particular to improve the system vs true
idle time accounting the arch code needs to measure the true idle time
instead of just testing for the idle process.
To improve the tick based accounting as well we would need an architecture
primitive that can tell us if the pt_regs of the interrupted context
points to the magic instruction that halts the cpu.
In addition idle time is no more added to the stime of the idle process.
This field now contains the system time of the idle process as it should
be. On systems without VIRT_CPU_ACCOUNTING this will always be zero as
every tick that occurs while idle is running will be accounted as idle
time.
This patch contains the necessary common code changes to be able to
distinguish idle system time and true idle time. The architectures with
support for VIRT_CPU_ACCOUNTING need some changes to exploit this.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Impact: use new API, remove cpumask from stack.
Change smp_call_function_mask() callers to smp_call_function_many().
This removes a cpumask from the stack, and falls back should allocating
the cpumask var fail (only possible with CONFIG_CPUMASKS_OFFSTACK).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Cc: jeremy@xensource.com
This patch simply changes cpumask_t to struct cpumask and similar
trivial modernizations.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Simple change, and eventual space saving when NR_CPUS >> nr_cpu_ids.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Mike Travis <travis@sgi.com>
Cc: Jeremy Fitzhardinge <jeremy@xensource.com>
Impact: cleanup, change parameter passing
* Change genapic interfaces to accept cpumask_t pointers where possible.
* Modify external callers to use cpumask_t pointers in function calls.
* Create new send_IPI_mask_allbutself which is the same as the
send_IPI_mask functions but removes smp_processor_id() from list.
This removes another common need for a temporary cpumask_t variable.
* Functions that used a temp cpumask_t variable for:
cpumask_t allbutme = cpu_online_map;
cpu_clear(smp_processor_id(), allbutme);
if (!cpus_empty(allbutme))
...
become:
if (!cpus_equal(cpu_online_map, cpumask_of_cpu(cpu)))
...
* Other minor code optimizations (like using cpus_clear instead of
CPU_MASK_NONE, etc.)
Applies to linux-2.6.tip/master.
Signed-off-by: Mike Travis <travis@sgi.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Ingo Molnar <mingo@elte.hu>
Impact: cleanup
hypervisor.h had accumulated a lot of crud, including lots of spurious
#includes. Clean it all up, and go around fixing up everything else
accordingly.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: change calling convention of existing clock_event APIs
struct clock_event_timer's cpumask field gets changed to take pointer,
as does the ->broadcast function.
Another single-patch change. For safety, we BUG_ON() in
clockevents_register_device() if it's not set.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: Ingo Molnar <mingo@elte.hu>
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86: always define DECLARE_PCI_UNMAP* macros
x86: fixup config space size of CPU functions for AMD family 11h
x86, bts: fix wrmsr and spinlock over kmalloc
x86, pebs: fix PEBS record size configuration
x86, bts: turn macro into static inline function
x86, bts: exclude ds.c from build when disabled
arch/x86/kernel/pci-calgary_64.c: change simple_strtol to simple_strtoul
x86: use limited register constraint for setnz
xen: pin correct PGD on suspend
x86: revert irq number limitation
x86: fixing __cpuinit/__init tangle, xsave_cntxt_init()
x86: fix __cpuinit/__init tangle in init_thread_xstate()
oprofile: fix an overflow in ppro code
... so get xen-ops.h in agreement with xen/smp.c
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86, xen: fix use of pgd_page now that it really does return a page
Xen requires that all mappings of pagetable pages are read-only, so
that they can't be updated illegally. As a result, if a page is being
turned into a pagetable page, we need to make sure all its mappings
are RO.
If the page had been used for ioremap or vmalloc, it may still have
left over mappings as a result of not having been lazily unmapped.
This change makes sure we explicitly mop them all up before pinning
the page.
Unlike aliases created by kmap, the there can be vmalloc aliases even
for non-high pages, so we must do the flush unconditionally.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Linux Memory Management List <linux-mm@kvack.org>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Impact: fix 32-bit Xen guest boot crash
On 32-bit PAE, pud_page, for no good reason, didn't really return a
struct page *. Since Jan Beulich's fix "i386/PAE: fix pud_page()",
pud_page does return a struct page *.
Because PAE has 3 pagetable levels, the pud level is folded into the
pgd level, so pgd_page() is the same as pud_page(), and now returns
a struct page *. Update the xen/mmu.c code which uses pgd_page()
accordingly.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
* 'tracing-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (31 commits)
ftrace: fix current_tracer error return
tracing: fix a build error on alpha
ftrace: use a real variable for ftrace_nop in x86
tracing/ftrace: make boot tracer select the sched_switch tracer
tracepoint: check if the probe has been registered
asm-generic: define DIE_OOPS in asm-generic
trace: fix printk warning for u64
ftrace: warning in kernel/trace/ftrace.c
ftrace: fix build failure
ftrace, powerpc, sparc64, x86: remove notrace from arch ftrace file
ftrace: remove ftrace hash
ftrace: remove mcount set
ftrace: remove daemon
ftrace: disable dynamic ftrace for all archs that use daemon
ftrace: add ftrace warn on to disable ftrace
ftrace: only have ftrace_kill atomic
ftrace: use probe_kernel
ftrace: comment arch ftrace code
ftrace: return error on failed modified text.
ftrace: dynamic ftrace process only text section
...
Impact: fix guest kernel boot crash on certain configs
Recent i686 2.6.27 kernels with a certain amount of memory (between
736 and 855MB) have a problem booting under a hypervisor that supports
batched mprotect (this includes the RHEL-5 Xen hypervisor as well as
any 3.3 or later Xen hypervisor).
The problem ends up being that xen_ptep_modify_prot_commit() is using
virt_to_machine to calculate which pfn to update. However, this only
works for pages that are in the p2m list, and the pages coming from
change_pte_range() in mm/mprotect.c are kmap_atomic pages. Because of
this, we can run into the situation where the lookup in the p2m table
returns an INVALID_MFN, which we then try to pass to the hypervisor,
which then (correctly) denies the request to a totally bogus pfn.
The right thing to do is to use arbitrary_virt_to_machine, so that we
can be sure we are modifying the right pfn. This unfortunately
introduces a performance penalty because of a full page-table-walk,
but we can avoid that penalty for pages in the p2m list by checking if
virt_addr_valid is true, and if so, just doing the lookup in the p2m
table.
The attached patch implements this, and allows my 2.6.27 i686 based
guest with 768MB of memory to boot on a RHEL-5 hypervisor again.
Thanks to Jeremy for the suggestions about how to fix this particular
issue.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This merges branches irq/genirq, irq/sparseirq-v4, timers/hpet-percpu
and x86/uv.
The sparseirq branch is just preliminary groundwork: no sparse IRQs are
actually implemented by this tree anymore - just the new APIs are added
while keeping the old way intact as well (the new APIs map 1:1 to
irq_desc[]). The 'real' sparse IRQ support will then be a relatively
small patch ontop of this - with a v2.6.29 merge target.
* 'genirq-v28-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (178 commits)
genirq: improve include files
intr_remapping: fix typo
io_apic: make irq_mis_count available on 64-bit too
genirq: fix name space collisions of nr_irqs in arch/*
genirq: fix name space collision of nr_irqs in autoprobe.c
genirq: use iterators for irq_desc loops
proc: fixup irq iterator
genirq: add reverse iterator for irq_desc
x86: move ack_bad_irq() to irq.c
x86: unify show_interrupts() and proc helpers
x86: cleanup show_interrupts
genirq: cleanup the sparseirq modifications
genirq: remove artifacts from sparseirq removal
genirq: revert dynarray
genirq: remove irq_to_desc_alloc
genirq: remove sparse irq code
genirq: use inline function for irq_to_desc
genirq: consolidate nr_irqs and for_each_irq_desc()
x86: remove sparse irq from Kconfig
genirq: define nr_irqs for architectures with GENERIC_HARDIRQS=n
...
Due to confusion between the ftrace infrastructure and the gcc profiling
tracer "ftrace", this patch renames the config options from FTRACE to
FUNCTION_TRACER. The other two names that are offspring from FTRACE
DYNAMIC_FTRACE and FTRACE_MCOUNT_RECORD will stay the same.
This patch was generated mostly by script, and partially by hand.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
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>
Ingo Molnar