diff options
author | Paul E. McKenney <paulmck@linux.vnet.ibm.com> | 2008-12-18 21:55:32 +0100 |
---|---|---|
committer | Ingo Molnar <mingo@elte.hu> | 2008-12-18 21:56:04 +0100 |
commit | 64db4cfff99c04cd5f550357edcc8780f96b54a2 (patch) | |
tree | 4856e788d21f0e31ed78a22b70b4521f7237705e /include/linux/rcutree.h | |
parent | d110ec3a1e1f522e2e9dfceb9c36d6590c26d2d4 (diff) |
"Tree RCU": scalable classic RCU implementation
This patch fixes a long-standing performance bug in classic RCU that
results in massive internal-to-RCU lock contention on systems with
more than a few hundred CPUs. Although this patch creates a separate
flavor of RCU for ease of review and patch maintenance, it is intended
to replace classic RCU.
This patch still handles stress better than does mainline, so I am still
calling it ready for inclusion. This patch is against the -tip tree.
Nevertheless, experience on an actual 1000+ CPU machine would still be
most welcome.
Most of the changes noted below were found while creating an rcutiny
(which should permit ejecting the current rcuclassic) and while doing
detailed line-by-line documentation.
Updates from v9 (http://lkml.org/lkml/2008/12/2/334):
o Fixes from remainder of line-by-line code walkthrough,
including comment spelling, initialization, undesirable
narrowing due to type conversion, removing redundant memory
barriers, removing redundant local-variable initialization,
and removing redundant local variables.
I do not believe that any of these fixes address the CPU-hotplug
issues that Andi Kleen was seeing, but please do give it a whirl
in case the machine is smarter than I am.
A writeup from the walkthrough may be found at the following
URL, in case you are suffering from terminal insomnia or
masochism:
http://www.kernel.org/pub/linux/kernel/people/paulmck/tmp/rcutree-walkthrough.2008.12.16a.pdf
o Made rcutree tracing use seq_file, as suggested some time
ago by Lai Jiangshan.
o Added a .csv variant of the rcudata debugfs trace file, to allow
people having thousands of CPUs to drop the data into
a spreadsheet. Tested with oocalc and gnumeric. Updated
documentation to suit.
Updates from v8 (http://lkml.org/lkml/2008/11/15/139):
o Fix a theoretical race between grace-period initialization and
force_quiescent_state() that could occur if more than three
jiffies were required to carry out the grace-period
initialization. Which it might, if you had enough CPUs.
o Apply Ingo's printk-standardization patch.
o Substitute local variables for repeated accesses to global
variables.
o Fix comment misspellings and redundant (but harmless) increments
of ->n_rcu_pending (this latter after having explicitly added it).
o Apply checkpatch fixes.
Updates from v7 (http://lkml.org/lkml/2008/10/10/291):
o Fixed a number of problems noted by Gautham Shenoy, including
the cpu-stall-detection bug that he was having difficulty
convincing me was real. ;-)
o Changed cpu-stall detection to wait for ten seconds rather than
three in order to reduce false positive, as suggested by Ingo
Molnar.
o Produced a design document (http://lwn.net/Articles/305782/).
The act of writing this document uncovered a number of both
theoretical and "here and now" bugs as noted below.
o Fix dynticks_nesting accounting confusion, simplify WARN_ON()
condition, fix kerneldoc comments, and add memory barriers
in dynticks interface functions.
o Add more data to tracing.
o Remove unused "rcu_barrier" field from rcu_data structure.
o Count calls to rcu_pending() from scheduling-clock interrupt
to use as a surrogate timebase should jiffies stop counting.
o Fix a theoretical race between force_quiescent_state() and
grace-period initialization. Yes, initialization does have to
go on for some jiffies for this race to occur, but given enough
CPUs...
Updates from v6 (http://lkml.org/lkml/2008/9/23/448):
o Fix a number of checkpatch.pl complaints.
o Apply review comments from Ingo Molnar and Lai Jiangshan
on the stall-detection code.
o Fix several bugs in !CONFIG_SMP builds.
o Fix a misspelled config-parameter name so that RCU now announces
at boot time if stall detection is configured.
o Run tests on numerous combinations of configurations parameters,
which after the fixes above, now build and run correctly.
Updates from v5 (http://lkml.org/lkml/2008/9/15/92, bad subject line):
o Fix a compiler error in the !CONFIG_FANOUT_EXACT case (blew a
changeset some time ago, and finally got around to retesting
this option).
o Fix some tracing bugs in rcupreempt that caused incorrect
totals to be printed.
o I now test with a more brutal random-selection online/offline
script (attached). Probably more brutal than it needs to be
on the people reading it as well, but so it goes.
o A number of optimizations and usability improvements:
o Make rcu_pending() ignore the grace-period timeout when
there is no grace period in progress.
o Make force_quiescent_state() avoid going for a global
lock in the case where there is no grace period in
progress.
o Rearrange struct fields to improve struct layout.
o Make call_rcu() initiate a grace period if RCU was
idle, rather than waiting for the next scheduling
clock interrupt.
o Invoke rcu_irq_enter() and rcu_irq_exit() only when
idle, as suggested by Andi Kleen. I still don't
completely trust this change, and might back it out.
o Make CONFIG_RCU_TRACE be the single config variable
manipulated for all forms of RCU, instead of the prior
confusion.
o Document tracing files and formats for both rcupreempt
and rcutree.
Updates from v4 for those missing v5 given its bad subject line:
o Separated dynticks interface so that NMIs and irqs call separate
functions, greatly simplifying it. In particular, this code
no longer requires a proof of correctness. ;-)
o Separated dynticks state out into its own per-CPU structure,
avoiding the duplicated accounting.
o The case where a dynticks-idle CPU runs an irq handler that
invokes call_rcu() is now correctly handled, forcing that CPU
out of dynticks-idle mode.
o Review comments have been applied (thank you all!!!).
For but one example, fixed the dynticks-ordering issue that
Manfred pointed out, saving me much debugging. ;-)
o Adjusted rcuclassic and rcupreempt to handle dynticks changes.
Attached is an updated patch to Classic RCU that applies a hierarchy,
greatly reducing the contention on the top-level lock for large machines.
This passes 10-hour concurrent rcutorture and online-offline testing on
128-CPU ppc64 without dynticks enabled, and exposes some timekeeping
bugs in presence of dynticks (exciting working on a system where
"sleep 1" hangs until interrupted...), which were fixed in the
2.6.27 kernel. It is getting more reliable than mainline by some
measures, so the next version will be against -tip for inclusion.
See also Manfred Spraul's recent patches (or his earlier work from
2004 at http://marc.info/?l=linux-kernel&m=108546384711797&w=2).
We will converge onto a common patch in the fullness of time, but are
currently exploring different regions of the design space. That said,
I have already gratefully stolen quite a few of Manfred's ideas.
This patch provides CONFIG_RCU_FANOUT, which controls the bushiness
of the RCU hierarchy. Defaults to 32 on 32-bit machines and 64 on
64-bit machines. If CONFIG_NR_CPUS is less than CONFIG_RCU_FANOUT,
there is no hierarchy. By default, the RCU initialization code will
adjust CONFIG_RCU_FANOUT to balance the hierarchy, so strongly NUMA
architectures may choose to set CONFIG_RCU_FANOUT_EXACT to disable
this balancing, allowing the hierarchy to be exactly aligned to the
underlying hardware. Up to two levels of hierarchy are permitted
(in addition to the root node), allowing up to 16,384 CPUs on 32-bit
systems and up to 262,144 CPUs on 64-bit systems. I just know that I
am going to regret saying this, but this seems more than sufficient
for the foreseeable future. (Some architectures might wish to set
CONFIG_RCU_FANOUT=4, which would limit such architectures to 64 CPUs.
If this becomes a real problem, additional levels can be added, but I
doubt that it will make a significant difference on real hardware.)
In the common case, a given CPU will manipulate its private rcu_data
structure and the rcu_node structure that it shares with its immediate
neighbors. This can reduce both lock and memory contention by multiple
orders of magnitude, which should eliminate the need for the strange
manipulations that are reported to be required when running Linux on
very large systems.
Some shortcomings:
o More bugs will probably surface as a result of an ongoing
line-by-line code inspection.
Patches will be provided as required.
o There are probably hangs, rcutorture failures, &c. Seems
quite stable on a 128-CPU machine, but that is kind of small
compared to 4096 CPUs. However, seems to do better than
mainline.
Patches will be provided as required.
o The memory footprint of this version is several KB larger
than rcuclassic.
A separate UP-only rcutiny patch will be provided, which will
reduce the memory footprint significantly, even compared
to the old rcuclassic. One such patch passes light testing,
and has a memory footprint smaller even than rcuclassic.
Initial reaction from various embedded guys was "it is not
worth it", so am putting it aside.
Credits:
o Manfred Spraul for ideas, review comments, and bugs spotted,
as well as some good friendly competition. ;-)
o Josh Triplett, Ingo Molnar, Peter Zijlstra, Mathieu Desnoyers,
Lai Jiangshan, Andi Kleen, Andy Whitcroft, and Andrew Morton
for reviews and comments.
o Thomas Gleixner for much-needed help with some timer issues
(see patches below).
o Jon M. Tollefson, Tim Pepper, Andrew Theurer, Jose R. Santos,
Andy Whitcroft, Darrick Wong, Nishanth Aravamudan, Anton
Blanchard, Dave Kleikamp, and Nathan Lynch for keeping machines
alive despite my heavy abuse^Wtesting.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Diffstat (limited to 'include/linux/rcutree.h')
-rw-r--r-- | include/linux/rcutree.h | 329 |
1 files changed, 329 insertions, 0 deletions
diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h new file mode 100644 index 00000000000..d4368b7975c --- /dev/null +++ b/include/linux/rcutree.h @@ -0,0 +1,329 @@ +/* + * Read-Copy Update mechanism for mutual exclusion (tree-based version) + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright IBM Corporation, 2008 + * + * Author: Dipankar Sarma <dipankar@in.ibm.com> + * Paul E. McKenney <paulmck@linux.vnet.ibm.com> Hierarchical algorithm + * + * Based on the original work by Paul McKenney <paulmck@us.ibm.com> + * and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen. + * + * For detailed explanation of Read-Copy Update mechanism see - + * Documentation/RCU + */ + +#ifndef __LINUX_RCUTREE_H +#define __LINUX_RCUTREE_H + +#include <linux/cache.h> +#include <linux/spinlock.h> +#include <linux/threads.h> +#include <linux/percpu.h> +#include <linux/cpumask.h> +#include <linux/seqlock.h> + +/* + * Define shape of hierarchy based on NR_CPUS and CONFIG_RCU_FANOUT. + * In theory, it should be possible to add more levels straightforwardly. + * In practice, this has not been tested, so there is probably some + * bug somewhere. + */ +#define MAX_RCU_LVLS 3 +#define RCU_FANOUT (CONFIG_RCU_FANOUT) +#define RCU_FANOUT_SQ (RCU_FANOUT * RCU_FANOUT) +#define RCU_FANOUT_CUBE (RCU_FANOUT_SQ * RCU_FANOUT) + +#if NR_CPUS <= RCU_FANOUT +# define NUM_RCU_LVLS 1 +# define NUM_RCU_LVL_0 1 +# define NUM_RCU_LVL_1 (NR_CPUS) +# define NUM_RCU_LVL_2 0 +# define NUM_RCU_LVL_3 0 +#elif NR_CPUS <= RCU_FANOUT_SQ +# define NUM_RCU_LVLS 2 +# define NUM_RCU_LVL_0 1 +# define NUM_RCU_LVL_1 (((NR_CPUS) + RCU_FANOUT - 1) / RCU_FANOUT) +# define NUM_RCU_LVL_2 (NR_CPUS) +# define NUM_RCU_LVL_3 0 +#elif NR_CPUS <= RCU_FANOUT_CUBE +# define NUM_RCU_LVLS 3 +# define NUM_RCU_LVL_0 1 +# define NUM_RCU_LVL_1 (((NR_CPUS) + RCU_FANOUT_SQ - 1) / RCU_FANOUT_SQ) +# define NUM_RCU_LVL_2 (((NR_CPUS) + (RCU_FANOUT) - 1) / (RCU_FANOUT)) +# define NUM_RCU_LVL_3 NR_CPUS +#else +# error "CONFIG_RCU_FANOUT insufficient for NR_CPUS" +#endif /* #if (NR_CPUS) <= RCU_FANOUT */ + +#define RCU_SUM (NUM_RCU_LVL_0 + NUM_RCU_LVL_1 + NUM_RCU_LVL_2 + NUM_RCU_LVL_3) +#define NUM_RCU_NODES (RCU_SUM - NR_CPUS) + +/* + * Dynticks per-CPU state. + */ +struct rcu_dynticks { + int dynticks_nesting; /* Track nesting level, sort of. */ + int dynticks; /* Even value for dynticks-idle, else odd. */ + int dynticks_nmi; /* Even value for either dynticks-idle or */ + /* not in nmi handler, else odd. So this */ + /* remains even for nmi from irq handler. */ +}; + +/* + * Definition for node within the RCU grace-period-detection hierarchy. + */ +struct rcu_node { + spinlock_t lock; + unsigned long qsmask; /* CPUs or groups that need to switch in */ + /* order for current grace period to proceed.*/ + unsigned long qsmaskinit; + /* Per-GP initialization for qsmask. */ + unsigned long grpmask; /* Mask to apply to parent qsmask. */ + int grplo; /* lowest-numbered CPU or group here. */ + int grphi; /* highest-numbered CPU or group here. */ + u8 grpnum; /* CPU/group number for next level up. */ + u8 level; /* root is at level 0. */ + struct rcu_node *parent; +} ____cacheline_internodealigned_in_smp; + +/* Index values for nxttail array in struct rcu_data. */ +#define RCU_DONE_TAIL 0 /* Also RCU_WAIT head. */ +#define RCU_WAIT_TAIL 1 /* Also RCU_NEXT_READY head. */ +#define RCU_NEXT_READY_TAIL 2 /* Also RCU_NEXT head. */ +#define RCU_NEXT_TAIL 3 +#define RCU_NEXT_SIZE 4 + +/* Per-CPU data for read-copy update. */ +struct rcu_data { + /* 1) quiescent-state and grace-period handling : */ + long completed; /* Track rsp->completed gp number */ + /* in order to detect GP end. */ + long gpnum; /* Highest gp number that this CPU */ + /* is aware of having started. */ + long passed_quiesc_completed; + /* Value of completed at time of qs. */ + bool passed_quiesc; /* User-mode/idle loop etc. */ + bool qs_pending; /* Core waits for quiesc state. */ + bool beenonline; /* CPU online at least once. */ + struct rcu_node *mynode; /* This CPU's leaf of hierarchy */ + unsigned long grpmask; /* Mask to apply to leaf qsmask. */ + + /* 2) batch handling */ + /* + * If nxtlist is not NULL, it is partitioned as follows. + * Any of the partitions might be empty, in which case the + * pointer to that partition will be equal to the pointer for + * the following partition. When the list is empty, all of + * the nxttail elements point to nxtlist, which is NULL. + * + * [*nxttail[RCU_NEXT_READY_TAIL], NULL = *nxttail[RCU_NEXT_TAIL]): + * Entries that might have arrived after current GP ended + * [*nxttail[RCU_WAIT_TAIL], *nxttail[RCU_NEXT_READY_TAIL]): + * Entries known to have arrived before current GP ended + * [*nxttail[RCU_DONE_TAIL], *nxttail[RCU_WAIT_TAIL]): + * Entries that batch # <= ->completed - 1: waiting for current GP + * [nxtlist, *nxttail[RCU_DONE_TAIL]): + * Entries that batch # <= ->completed + * The grace period for these entries has completed, and + * the other grace-period-completed entries may be moved + * here temporarily in rcu_process_callbacks(). + */ + struct rcu_head *nxtlist; + struct rcu_head **nxttail[RCU_NEXT_SIZE]; + long qlen; /* # of queued callbacks */ + long blimit; /* Upper limit on a processed batch */ + +#ifdef CONFIG_NO_HZ + /* 3) dynticks interface. */ + struct rcu_dynticks *dynticks; /* Shared per-CPU dynticks state. */ + int dynticks_snap; /* Per-GP tracking for dynticks. */ + int dynticks_nmi_snap; /* Per-GP tracking for dynticks_nmi. */ +#endif /* #ifdef CONFIG_NO_HZ */ + + /* 4) reasons this CPU needed to be kicked by force_quiescent_state */ +#ifdef CONFIG_NO_HZ + unsigned long dynticks_fqs; /* Kicked due to dynticks idle. */ +#endif /* #ifdef CONFIG_NO_HZ */ + unsigned long offline_fqs; /* Kicked due to being offline. */ + unsigned long resched_ipi; /* Sent a resched IPI. */ + + /* 5) state to allow this CPU to force_quiescent_state on others */ + long n_rcu_pending; /* rcu_pending() calls since boot. */ + long n_rcu_pending_force_qs; /* when to force quiescent states. */ + + int cpu; +}; + +/* Values for signaled field in struct rcu_state. */ +#define RCU_GP_INIT 0 /* Grace period being initialized. */ +#define RCU_SAVE_DYNTICK 1 /* Need to scan dyntick state. */ +#define RCU_FORCE_QS 2 /* Need to force quiescent state. */ +#ifdef CONFIG_NO_HZ +#define RCU_SIGNAL_INIT RCU_SAVE_DYNTICK +#else /* #ifdef CONFIG_NO_HZ */ +#define RCU_SIGNAL_INIT RCU_FORCE_QS +#endif /* #else #ifdef CONFIG_NO_HZ */ + +#define RCU_JIFFIES_TILL_FORCE_QS 3 /* for rsp->jiffies_force_qs */ +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR +#define RCU_SECONDS_TILL_STALL_CHECK (10 * HZ) /* for rsp->jiffies_stall */ +#define RCU_SECONDS_TILL_STALL_RECHECK (30 * HZ) /* for rsp->jiffies_stall */ +#define RCU_STALL_RAT_DELAY 2 /* Allow other CPUs time */ + /* to take at least one */ + /* scheduling clock irq */ + /* before ratting on them. */ + +#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ + +/* + * RCU global state, including node hierarchy. This hierarchy is + * represented in "heap" form in a dense array. The root (first level) + * of the hierarchy is in ->node[0] (referenced by ->level[0]), the second + * level in ->node[1] through ->node[m] (->node[1] referenced by ->level[1]), + * and the third level in ->node[m+1] and following (->node[m+1] referenced + * by ->level[2]). The number of levels is determined by the number of + * CPUs and by CONFIG_RCU_FANOUT. Small systems will have a "hierarchy" + * consisting of a single rcu_node. + */ +struct rcu_state { + struct rcu_node node[NUM_RCU_NODES]; /* Hierarchy. */ + struct rcu_node *level[NUM_RCU_LVLS]; /* Hierarchy levels. */ + u32 levelcnt[MAX_RCU_LVLS + 1]; /* # nodes in each level. */ + u8 levelspread[NUM_RCU_LVLS]; /* kids/node in each level. */ + struct rcu_data *rda[NR_CPUS]; /* array of rdp pointers. */ + + /* The following fields are guarded by the root rcu_node's lock. */ + + u8 signaled ____cacheline_internodealigned_in_smp; + /* Force QS state. */ + long gpnum; /* Current gp number. */ + long completed; /* # of last completed gp. */ + spinlock_t onofflock; /* exclude on/offline and */ + /* starting new GP. */ + spinlock_t fqslock; /* Only one task forcing */ + /* quiescent states. */ + unsigned long jiffies_force_qs; /* Time at which to invoke */ + /* force_quiescent_state(). */ + unsigned long n_force_qs; /* Number of calls to */ + /* force_quiescent_state(). */ + unsigned long n_force_qs_lh; /* ~Number of calls leaving */ + /* due to lock unavailable. */ + unsigned long n_force_qs_ngp; /* Number of calls leaving */ + /* due to no GP active. */ +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR + unsigned long gp_start; /* Time at which GP started, */ + /* but in jiffies. */ + unsigned long jiffies_stall; /* Time at which to check */ + /* for CPU stalls. */ +#endif /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ +#ifdef CONFIG_NO_HZ + long dynticks_completed; /* Value of completed @ snap. */ +#endif /* #ifdef CONFIG_NO_HZ */ +}; + +extern struct rcu_state rcu_state; +DECLARE_PER_CPU(struct rcu_data, rcu_data); + +extern struct rcu_state rcu_bh_state; +DECLARE_PER_CPU(struct rcu_data, rcu_bh_data); + +/* + * Increment the quiescent state counter. + * The counter is a bit degenerated: We do not need to know + * how many quiescent states passed, just if there was at least + * one since the start of the grace period. Thus just a flag. + */ +static inline void rcu_qsctr_inc(int cpu) +{ + struct rcu_data *rdp = &per_cpu(rcu_data, cpu); + rdp->passed_quiesc = 1; + rdp->passed_quiesc_completed = rdp->completed; +} +static inline void rcu_bh_qsctr_inc(int cpu) +{ + struct rcu_data *rdp = &per_cpu(rcu_bh_data, cpu); + rdp->passed_quiesc = 1; + rdp->passed_quiesc_completed = rdp->completed; +} + +extern int rcu_pending(int cpu); +extern int rcu_needs_cpu(int cpu); + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +extern struct lockdep_map rcu_lock_map; +# define rcu_read_acquire() \ + lock_acquire(&rcu_lock_map, 0, 0, 2, 1, NULL, _THIS_IP_) +# define rcu_read_release() lock_release(&rcu_lock_map, 1, _THIS_IP_) +#else +# define rcu_read_acquire() do { } while (0) +# define rcu_read_release() do { } while (0) +#endif + +static inline void __rcu_read_lock(void) +{ + preempt_disable(); + __acquire(RCU); + rcu_read_acquire(); +} +static inline void __rcu_read_unlock(void) +{ + rcu_read_release(); + __release(RCU); + preempt_enable(); +} +static inline void __rcu_read_lock_bh(void) +{ + local_bh_disable(); + __acquire(RCU_BH); + rcu_read_acquire(); +} +static inline void __rcu_read_unlock_bh(void) +{ + rcu_read_release(); + __release(RCU_BH); + local_bh_enable(); +} + +#define __synchronize_sched() synchronize_rcu() + +#define call_rcu_sched(head, func) call_rcu(head, func) + +static inline void rcu_init_sched(void) +{ +} + +extern void __rcu_init(void); +extern void rcu_check_callbacks(int cpu, int user); +extern void rcu_restart_cpu(int cpu); + +extern long rcu_batches_completed(void); +extern long rcu_batches_completed_bh(void); + +#ifdef CONFIG_NO_HZ +void rcu_enter_nohz(void); +void rcu_exit_nohz(void); +#else /* CONFIG_NO_HZ */ +static inline void rcu_enter_nohz(void) +{ +} +static inline void rcu_exit_nohz(void) +{ +} +#endif /* CONFIG_NO_HZ */ + +#endif /* __LINUX_RCUTREE_H */ |