01 Nov, 2011
1 commit
-
These variables are only used when CONFIG_HOTPLUG_CPU is enabled, they are
ifdef'ed everywhere else. So don't define them when CONFIG_HOTPLUG_CPU is
not enabled.Signed-off-by: Glauber Costa
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
13 Sep, 2011
1 commit
-
The percpu_counter::lock can be taken in atomic context and therefore
cannot be preempted on -rt - annotate it.In mainline this change documents the low level nature of
the lock - otherwise there's no functional difference. Lockdep
and Sparse checking will work as usual.Signed-off-by: Thomas Gleixner
Signed-off-by: Ingo Molnar
17 Dec, 2010
1 commit
-
The this_cpu_* options can be used to optimize __percpu_counter_add a bit. Avoids
some address arithmetic and saves 12 bytes.Before:
00000000000001d3 :
1d3: 55 push %rbp
1d4: 48 89 e5 mov %rsp,%rbp
1d7: 41 55 push %r13
1d9: 41 54 push %r12
1db: 53 push %rbx
1dc: 48 89 fb mov %rdi,%rbx
1df: 48 83 ec 08 sub $0x8,%rsp
1e3: 4c 8b 67 30 mov 0x30(%rdi),%r12
1e7: 65 4c 03 24 25 00 00 add %gs:0x0,%r12
1ee: 00 00
1f0: 4d 63 2c 24 movslq (%r12),%r13
1f4: 48 63 c2 movslq %edx,%rax
1f7: 49 01 f5 add %rsi,%r13
1fa: 49 39 c5 cmp %rax,%r13
1fd: 7d 0a jge 209
1ff: f7 da neg %edx
201: 48 63 d2 movslq %edx,%rdx
204: 49 39 d5 cmp %rdx,%r13
207: 7f 1e jg 227
209: 48 89 df mov %rbx,%rdi
20c: e8 00 00 00 00 callq 211
211: 4c 01 6b 18 add %r13,0x18(%rbx)
215: 48 89 df mov %rbx,%rdi
218: 41 c7 04 24 00 00 00 movl $0x0,(%r12)
21f: 00
220: e8 00 00 00 00 callq 225
225: eb 04 jmp 22b
227: 45 89 2c 24 mov %r13d,(%r12)
22b: 5b pop %rbx
22c: 5b pop %rbx
22d: 41 5c pop %r12
22f: 41 5d pop %r13
231: c9 leaveq
232: c3 retqAfter:
00000000000001d3 :
1d3: 55 push %rbp
1d4: 48 63 ca movslq %edx,%rcx
1d7: 48 89 e5 mov %rsp,%rbp
1da: 41 54 push %r12
1dc: 53 push %rbx
1dd: 48 89 fb mov %rdi,%rbx
1e0: 48 8b 47 30 mov 0x30(%rdi),%rax
1e4: 65 44 8b 20 mov %gs:(%rax),%r12d
1e8: 4d 63 e4 movslq %r12d,%r12
1eb: 49 01 f4 add %rsi,%r12
1ee: 49 39 cc cmp %rcx,%r12
1f1: 7d 0a jge 1fd
1f3: f7 da neg %edx
1f5: 48 63 d2 movslq %edx,%rdx
1f8: 49 39 d4 cmp %rdx,%r12
1fb: 7f 21 jg 21e
1fd: 48 89 df mov %rbx,%rdi
200: e8 00 00 00 00 callq 205
205: 4c 01 63 18 add %r12,0x18(%rbx)
209: 48 8b 43 30 mov 0x30(%rbx),%rax
20d: 48 89 df mov %rbx,%rdi
210: 65 c7 00 00 00 00 00 movl $0x0,%gs:(%rax)
217: e8 00 00 00 00 callq 21c
21c: eb 04 jmp 222
21e: 65 44 89 20 mov %r12d,%gs:(%rax)
222: 5b pop %rbx
223: 41 5c pop %r12
225: c9 leaveq
226: c3 retqReviewed-by: Pekka Enberg
Reviewed-by: Tejun Heo
Reviewed-by: Mathieu Desnoyers
Acked-by: H. Peter Anvin
Signed-off-by: Christoph Lameter
Signed-off-by: Tejun Heo
27 Oct, 2010
3 commits
-
this_cpu_ptr() avoids an array lookup and can use the percpu offset of the
local cpu directly.Signed-off-by: Christoph Lameter
Cc: Eric Dumazet
Cc: Tejun Heo
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
All percpu counters are linked to a global list on initialization and
removed from it on destruction. The list is walked during CPU up/down.
If a percpu counter is freed without being properly destroyed, the system
will oops only on the next CPU up/down making it pretty nasty to track
down. This patch adds debugobj support for percpu counters so that such
problems can be found easily.As percpu counters don't make sense on stack and can't be statically
initialized, debugobj support is pretty simple. It's initialized and
activated on counter initialization, and deactivatd and destroyed on
counter destruction. With this patch applied, the bug fixed by commit
602586a83b719df0fbd94196a1359ed35aeb2df3 (shmem: put_super must
percpu_counter_destroy) triggers the following warning on tmpfs unmount
and the system won't oops on the next cpu up/down operation.------------[ cut here ]------------
WARNING: at lib/debugobjects.c:259 debug_print_object+0x5c/0x70()
Hardware name: Bochs
ODEBUG: free active (active state 0) object type: percpu_counter
Modules linked in:
Pid: 3999, comm: umount Not tainted 2.6.36-rc2-work+ #5
Call Trace:
[] warn_slowpath_common+0x7f/0xc0
[] warn_slowpath_fmt+0x46/0x50
[] debug_print_object+0x5c/0x70
[] debug_check_no_obj_freed+0x125/0x210
[] kfree+0xb3/0x2f0
[] shmem_put_super+0x1d/0x30
[] generic_shutdown_super+0x56/0xe0
[] kill_anon_super+0x16/0x60
[] kill_litter_super+0x27/0x30
[] deactivate_locked_super+0x45/0x60
[] deactivate_super+0x4a/0x70
[] mntput_no_expire+0x86/0xe0
[] sys_umount+0x6f/0x360
[] system_call_fastpath+0x16/0x1b
---[ end trace cce2a341ba3611a7 ]---Signed-off-by: Tejun Heo
Acked-by: Thomas Gleixner
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
WARNING: at lib/list_debug.c:26 __list_add+0x3f/0x81()
Hardware name: Express5800/B120a [N8400-085]
list_add corruption. next->prev should be prev (ffffffff81a7ea00), but was dead000000200200. (next=ffff88080b872d58).
Modules linked in: aoe ipt_MASQUERADE iptable_nat nf_nat autofs4 sunrpc bridge 8021q garp stp llc ipv6 cpufreq_ondemand acpi_cpufreq freq_table dm_round_robin dm_multipath kvm_intel kvm uinput lpfc scsi_transport_fc igb ioatdma scsi_tgt i2c_i801 i2c_core dca iTCO_wdt iTCO_vendor_support pcspkr shpchp megaraid_sas [last unloaded: aoe]
Pid: 54, comm: events/3 Tainted: G W 2.6.34-vanilla1 #1
Call Trace:
[] warn_slowpath_common+0x7c/0x94
[] warn_slowpath_fmt+0x41/0x43
[] __list_add+0x3f/0x81
[] __percpu_counter_init+0x59/0x6b
[] bdi_init+0x118/0x17e
[] blk_alloc_queue_node+0x79/0x143
[] blk_alloc_queue+0x11/0x13
[] aoeblk_gdalloc+0x8e/0x1c9 [aoe]
[] aoecmd_sleepwork+0x25/0xa8 [aoe]
[] worker_thread+0x1a9/0x237
[] ? aoecmd_sleepwork+0x0/0xa8 [aoe]
[] ? autoremove_wake_function+0x0/0x39
[] ? worker_thread+0x0/0x237
[] kthread+0x7f/0x87
[] kernel_thread_helper+0x4/0x10
[] ? kthread+0x0/0x87
[] ? kernel_thread_helper+0x0/0x10It's because there is no initialization code for a list_head contained in
the struct backing_dev_info under CONFIG_HOTPLUG_CPU, and the bug comes up
when block device drivers calling blk_alloc_queue() are used. In case of
me, I got them by using aoe.Signed-off-by: Masanori Itoh
Cc: Tejun Heo
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
10 Aug, 2010
1 commit
-
Add percpu_counter_compare that allows for a quick but accurate comparison
of percpu_counter with a given value.A rough count is provided by the count field in percpu_counter structure,
without accounting for the other values stored in individual cpu counters.The actual count is a sum of count and the cpu counters. However, count
field is never different from the actual value by a factor of
batch*num_online_cpu. We do not need to get actual count for comparison
if count is different from the given value by this factor and allows for
quick comparison without summing up all the per cpu counters.Signed-off-by: Tim Chen
Cc: Hugh Dickins
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
07 Jan, 2009
2 commits
-
…/git/tip/linux-2.6-tip
* 'core-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
rcu: fix rcutorture bug
rcu: eliminate synchronize_rcu_xxx macro
rcu: make treercu safe for suspend and resume
rcu: fix rcutree grace-period-latency bug on small systems
futex: catch certain assymetric (get|put)_futex_key calls
futex: make futex_(get|put)_key() calls symmetric
locking, percpu counters: introduce separate lock classes
swiotlb: clean up EXPORT_SYMBOL usage
swiotlb: remove unnecessary declaration
swiotlb: replace architecture-specific swiotlb.h with linux/swiotlb.h
swiotlb: add support for systems with highmem
swiotlb: store phys address in io_tlb_orig_addr array
swiotlb: add hwdev to swiotlb_phys_to_bus() / swiotlb_sg_to_bus() -
For NR_CPUS >= 16 values, FBC_BATCH is 2*NR_CPUS
Considering more and more distros are using high NR_CPUS values, it makes
sense to use a more sensible value for FBC_BATCH, and get rid of NR_CPUS.A sensible value is 2*num_online_cpus(), with a minimum value of 32 (This
minimum value helps branch prediction in __percpu_counter_add())We already have a hotcpu notifier, so we can adjust FBC_BATCH dynamically.
We rename FBC_BATCH to percpu_counter_batch since its not a constant
anymore.Signed-off-by: Eric Dumazet
Acked-by: David S. Miller
Acked-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
06 Jan, 2009
1 commit
29 Dec, 2008
1 commit
-
Impact: fix lockdep false positives
Classify percpu_counter instances similar to regular lock objects --
that is, per instantiation site.The networking code has increased its use of percpu_counters, which
leads to false positives if they are treated as a single class.Signed-off-by: Peter Zijlstra
Signed-off-by: Ingo Molnar
11 Dec, 2008
3 commits
-
Revert
commit e8ced39d5e8911c662d4d69a342b9d053eaaac4e
Author: Mingming Cao
Date: Fri Jul 11 19:27:31 2008 -0400percpu_counter: new function percpu_counter_sum_and_set
As described in
revert "percpu counter: clean up percpu_counter_sum_and_set()"
the new percpu_counter_sum_and_set() is racy against updates to the
cpu-local accumulators on other CPUs. Revert that change.This means that ext4 will be slow again. But correct.
Reported-by: Eric Dumazet
Cc: "David S. Miller"
Cc: Peter Zijlstra
Cc: Mingming Cao
Cc:
Cc: [2.6.27.x]
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Revert
commit 1f7c14c62ce63805f9574664a6c6de3633d4a354
Author: Mingming Cao
Date: Thu Oct 9 12:50:59 2008 -0400percpu counter: clean up percpu_counter_sum_and_set()
Before this patch we had the following:
percpu_counter_sum(): return the percpu_counter's value
percpu_counter_sum_and_set(): return the percpu_counter's value, copying
that value into the central value and zeroing the per-cpu counters before
returning.After this patch, percpu_counter_sum_and_set() has gone, and
percpu_counter_sum() gets the old percpu_counter_sum_and_set()
functionality.Problem is, as Eric points out, the old percpu_counter_sum_and_set()
functionality was racy and wrong. It zeroes out counters on "other" cpus,
without holding any locks which will prevent races agaist updates from
those other CPUS.This patch reverts 1f7c14c62ce63805f9574664a6c6de3633d4a354. This means
that percpu_counter_sum_and_set() still has the race, but
percpu_counter_sum() does not.Note that this is not a simple revert - ext4 has since started using
percpu_counter_sum() for its dirty_blocks counter as well.Note that this revert patch changes percpu_counter_sum() semantics.
Before the patch, a call to percpu_counter_sum() will bring the counter's
central counter mostly up-to-date, so a following percpu_counter_read()
will return a close value.After this patch, a call to percpu_counter_sum() will leave the counter's
central accumulator unaltered, so a subsequent call to
percpu_counter_read() can now return a significantly inaccurate result.If there is any code in the tree which was introduced after
e8ced39d5e8911c662d4d69a342b9d053eaaac4e was merged, and which depends
upon the new percpu_counter_sum() semantics, that code will break.Reported-by: Eric Dumazet
Cc: "David S. Miller"
Cc: Peter Zijlstra
Cc: Mingming Cao
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
We should first delete the counter from percpu_counters list
before freeing memory, or a percpu_counter_hotcpu_callback()
could dereference a NULL pointer.Signed-off-by: Eric Dumazet
Acked-by: David S. Miller
Cc: Peter Zijlstra
Cc: Mingming Cao
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
10 Oct, 2008
1 commit
-
percpu_counter_sum_and_set() and percpu_counter_sum() is the same except
the former updates the global counter after accounting. Since we are
taking the fbc->lock to calculate the precise value of the counter in
percpu_counter_sum() anyway, it should simply set fbc->count too, as the
percpu_counter_sum_and_set() does.This patch merges these two interfaces into one.
Signed-off-by: Mingming Cao
Acked-by: Peter Zijlstra
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: "Theodore Ts'o"
12 Jul, 2008
1 commit
-
Delayed allocation need to check free blocks at every write time.
percpu_counter_read_positive() is not quit accurate. delayed
allocation need a more accurate accounting, but using
percpu_counter_sum_positive() is frequently is quite expensive.This patch added a new function to update center counter when sum
per-cpu counter, to increase the accurate rate for next
percpu_counter_read() and require less calling expensive
percpu_counter_sum().Signed-off-by: Mingming Cao
Signed-off-by: "Theodore Ts'o"
30 Apr, 2008
1 commit
-
Provide a place in sysfs (/sys/class/bdi) for the backing_dev_info object.
This allows us to see and set the various BDI specific variables.In particular this properly exposes the read-ahead window for all relevant
users and /sys/block//queue/read_ahead_kb should be deprecated.With patient help from Kay Sievers and Greg KH
[mszeredi@suse.cz]
- split off NFS and FUSE changes into separate patches
- document new sysfs attributes under Documentation/ABI
- do bdi_class_init as a core_initcall, otherwise the "default" BDI
won't be initialized
- remove bdi_init_fmt macro, it's not used very much[akpm@linux-foundation.org: fix ia64 warning]
Signed-off-by: Peter Zijlstra
Cc: Kay Sievers
Acked-by: Greg KH
Cc: Trond Myklebust
Signed-off-by: Miklos Szeredi
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
20 Oct, 2007
1 commit
-
Some of the per-cpu counters and thus their locks are accessed from IRQ
contexts. This can cause a deadlock if it interrupts a cpu-offline thread
which is transferring a dead-cpu's counts to the global counter.Add appropriate IRQ protection in the cpu-hotplug callback path.
Signed-off-by: Gautham R Shenoy
Cc: KAMEZAWA Hiroyuki
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
17 Oct, 2007
8 commits
-
provide a way to tell lockdep about percpu_counters that are supposed to be
used from irq safe contexts.Signed-off-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
alloc_percpu can fail, propagate that error.
Signed-off-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Provide an accurate version of percpu_counter_read.
Signed-off-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
s/percpu_counter_sum/&_positive/
Because its consitent with percpu_counter_read*
Signed-off-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Provide a method to set a percpu counter to a specified value.
Signed-off-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
percpu_counter is a s64 counter, make _add consitent.
Signed-off-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Because the current batch setup has an quadric error bound on the counter,
allow for an alternative setup.Signed-off-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
s/percpu_counter_mod/percpu_counter_add/
Because its a better name, _mod implies modulo.
Signed-off-by: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
17 Jul, 2007
2 commits
-
Now that we have implemented hotunplug-time counter spilling,
percpu_counter_sum() only needs to look at online CPUs.Cc: Gautham R Shenoy
Cc: Oleg Nesterov
Cc: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
per-cpu counters presently must iterate over all possible CPUs in the
exhaustive percpu_counter_sum().But it can be much better to only iterate over the presently-online CPUs. To
do this, we must arrange for an offlined CPU's count to be spilled into the
counter's central count.We can do this for all percpu_counters in the machine by linking them into a
single global list and walking that list at CPU_DEAD time.(I hope. Might have race windows in which the percpu_counter_sum() count is
inaccurate?)Cc: Gautham R Shenoy
Cc: Oleg Nesterov
Cc: Peter Zijlstra
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
23 Jun, 2006
2 commits
-
The percpu counter data type are changed in this set of patches to support
more users like ext3 who need more than 32 bit to store the free blocks
total in the filesystem.- Generic perpcu counters data type changes. The size of the global counter
and local counter were explictly specified using s64 and s32. The global
counter is changed from long to s64, while the local counter is changed from
long to s32, so we could avoid doing 64 bit update in most cases.- Users of the percpu counters are updated to make use of the new
percpu_counter_init() routine now taking an additional parameter to allow
users to pass the initial value of the global counter.Signed-off-by: Mingming Cao
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
- Move percpu_counter routines from mm/swap.c to lib/percpu_counter.c
Signed-off-by: Ravikiran Thirumalai
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
