login
Search
Hello! This patch fixes a long-standing performance bug in classic RCU that results in massive lock contention on the internal RCU lock on systems with more than a few hundred CPUs. Although this patch creates a separate flavor of RCU for easy of review and patch maintenance, it is intended to replace classic RCU. Still experimental, not for inclusion, but given that I am now finding more bugs in the rest of Linux than in this code, I suspect that it is getting close. Definitely ready for serious experimental use, especially in !CONFIG_NO_HZ configurations. In particular, experience on an actual 1000+ CPU machine would be most welcome, and now appears to be forthcoming! 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...). It is OK for experimental work, but not yet ready 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 There is a bit of debug code in place. This will be removed. o There is still a little strangeness in timekeeping when running a heavy dyntick and CPU hotplug workload concurrently with rcutorture, but there no longer seem to be cases of the jiffies counter deciding to stop counting (thank you, Thomas!!!). But I can make -that- happen -without- this patch. o There are probably hangs, rcutorture failures, &c. Seems reasonably stable on a 128-CPU machine, but that is kind of small compared to 4096. o There is not yet a human-readable design document. One is now in the works (finally!). 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, and Andrew Morton for reviews and comments. o Thomas Gleixner for much-needed help with some timer issues (see patches below). To build, start with 2.6.27-rc7, and apply: http://www.rdrop.com/users/paulmck/patches/2.6.27-rc3-treeRCU-20.patch http://tglx.de/~tglx/gack.patch http://tglx.de/~tglx/clockevents-keep-tick-next-period-up-to-date.patch Thoughts? Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> --- Documentation/RCU/00-INDEX | 2 Documentation/RCU/trace.txt | 389 ++++++++ arch/powerpc/platforms/pseries/rtasd.c | 4 include/linux/hardirq.h | 14 include/linux/rcupdate.h | 10 include/linux/rcutree.h | 320 +++++++ init/Kconfig | 18 kernel/Kconfig.preempt | 62 + kernel/Makefile | 6 kernel/rcupreempt.c | 10 kernel/rcupreempt_trace.c | 10 kernel/rcutree.c | 1502 +++++++++++++++++++++++++++++++++ kernel/rcutree_trace.c | 235 +++++ kernel/softirq.c | 15 lib/Kconfig.debug | 13 15 files changed, 2576 insertions(+), 34 deletions(-) diff --git a/Documentation/RCU/00-INDEX b/Documentation/RCU/00-INDEX index 461481d..7dc0695 100644 --- a/Documentation/RCU/00-INDEX +++ b/Documentation/RCU/00-INDEX @@ -16,6 +16,8 @@ RTFP.txt - List of RCU papers (bibliography) going back to 1980. torture.txt - RCU Torture Test Operation (CONFIG_RCU_TORTURE_TEST) +trace.txt + - CONFIG_RCU_TRACE debugfs files and formats UP.txt - RCU on Uniprocessor Systems whatisRCU.txt diff --git a/Documentation/RCU/trace.txt b/Documentation/RCU/trace.txt new file mode 100644 index 0000000..1357b97 --- /dev/null +++ b/Documentation/RCU/trace.txt @@ -0,0 +1,389 @@ +CONFIG_RCU_TRACE debugfs Files and Formats + + +The rcupreempt and rcutree implementations of RCU provide debugfs trace +output that summarizes counters and state. This information is useful for +debugging RCU itself, and can sometimes also help to debug abuses of RCU. +Note that the rcuclassic implementation of RCU does not provide debugfs +trace output. + +The following sections describe the debugfs files and formats for +preemptable RCU (rcupreempt) and hierarchical RCU (rcutree). + + +Preemptable RCU debugfs Files and Formats + +This implementation of RCU provides three debugfs files under the +top-level directory RCU: rcu/rcuctrs (which displays the per-CPU +counters used by preemptable RCU) rcu/rcugp (which displays grace-period +counters), and rcu/rcustats (which internal counters for debugging RCU). + +The output of "cat rcu/rcuctrs" looks as follows: + +CPU last cur F M + 0 5 -5 0 0 + 1 -1 0 0 0 + 2 0 1 0 0 + 3 0 1 0 0 + 4 0 1 0 0 + 5 0 1 0 0 + 6 0 2 0 0 + 7 0 -1 0 0 + 8 0 1 0 0 +ggp = 26226, state = waitzero + +The per-CPU fields are as follows: + +o "CPU" gives the CPU number. Offline CPUs are not displayed. + +o "last" gives the value of the counter that is being decremented + for the current grace period phase. In the example above, + the counters sum to 4, indicating that there are still four + RCU read-side critical sections still running that started + before the last counter flip. + +o "cur" gives the value of the counter that is currently being + both incremented (by rcu_read_lock()) and decremented (by + rcu_read_unlock()). In the example above, the counters sum to + 1, indicating that there is only one RCU read-side critical section + still running that started after the last counter flip. + +o "F" indicates whether RCU is waiting for this CPU to acknowledge + a counter flip. In the above example, RCU is not waiting on any, + which is consistent with the state being "waitzero" rather than + "waitack". + +o "M" indicates whether RCU is waiting for this CPU to execute a + memory barrier. In the above example, RCU is not waiting on any, + which is consistent with the state being "waitzero" rather than + "waitmb". + +o "ggp" is the global grace-period counter. + +o "state" is the RCU state, which can be one of the following: + + o "idle": there is no grace period in progress. + + o "waitack": RCU just incremented the global grace-period + counter, which has the effect of reversing the roles of + the "last" and "cur" counters above, and is waiting for + all the CPUs to acknowledge the flip. Once the flip has + been acknowledged, CPUs will no longer be incrementing + what are now the "last" counters, so that their sum will + decrease monotonically down to zero. + + o "waitzero": RCU is waiting for the sum of the "last" counters + to decrease to zero. + + o "waitmb": RCU is waiting for each CPU to execute a memory + barrier, which ensures that instructions from a given CPU's + last RCU read-side critical section cannot be reordered + with instructions following the memory-barrier instruction. + +The output of "cat rcu/rcugp" looks as follows: + +oldggp=48870 newggp=48873 + +Note that reading from this file provokes a synchronize_rcu(). The +"oldggp" value is that of "ggp" from rcu/rcuctrs above, taken before +executing the synchronize_rcu(), and the "newggp" value is also the +"ggp" value, but taken after the synchronize_rcu() command returns. + + +The output of "cat rcu/rcugp" looks as follows: + +na=1337955 nl=40 wa=1337915 wl=44 da=1337871 dl=0 dr=1337871 di=1337871 +1=50989 e1=6138 i1=49722 ie1=82 g1=49640 a1=315203 ae1=265563 a2=49640 +z1=1401244 ze1=1351605 z2=49639 m1=5661253 me1=5611614 m2=49639 + +These are counters tracking internal preemptable-RCU events, however, +some of them may be useful for debugging algorithms using RCU. In +particular, the "nl", "wl", and "dl" values track the number of RCU +callbacks in various states. The fields are as follows: + +o "na" is the total number of RCU callbacks that have been enqueued + since boot. + +o "nl" is the number of RCU callbacks waiting for the previous + grace period to end so that they can start waiting on the next + grace period. + +o "wa" is the total number of RCU callbacks that have started waiting + for a grace period since boot. "na" should be roughly equal to + "nl" plus "wa". + +o "wl" is the number of RCU callbacks currently waiting for their + grace period to end. + +o "da" is the total number of RCU callbacks whose grace periods + have completed since boot. "wa" should be roughly equal to + "wl" plus "da". + +o "di" is the total number of RCU callbacks that have been invoked + since boot. "di" should be roughly equal to "da", though some + early versions of preemptable RCU had a bug so that only the + last CPU's count of invocations was displayed, rather than the + sum of all CPU's counts. + +o "1" is the number of calls to rcu_try_flip(). + +o "e1" is the number of times that rcu_try_flip() was unable to + acquire the fliplock. + +o "i1" is the number of calls to rcu_try_flip_idle(). + +o "ie1" is the number of times rcu_try_flip_idle() exited early + due to the calling CPU having no work for RCU. + +o "g1" is the number of times that rcu_try_flip_idle() decided + to start a new grace period. "i1" should be roughly equal to + "ie1" plus "g1". + +o "a1" is the number of calls to rcu_try_flip_waitack(). + +o "ae1" is the number of times that rcu_try_flip_waitack() found + that at least one CPU had not yet acknowledge the new grace period + (AKA "counter flip"). + +o "a2" is the number of time rcu_try_flip_waitack() found that + all CPUs had acknowledged. "a1" should be roughly equal to + "ae1" plus "a2". (This particular output was collected on + a 128-CPU machine, hence the smaller-than-usual fraction of + calls to rcu_try_flip_waitack() finding all CPUs having already + acknowledged.) + +o "z1" is the number of calls to rcu_try_flip_waitzero(). + +o "ze1" is the number of times that rcu_try_flip_waitzero() found + that not all of the old RCU read-side critical sections had + completed. + +o "z2" is the number of times that rcu_try_flip_waitzero() finds + the sum of the counters equal to zero, in other words, that + all of the old RCU read-side critical sections had completed. + The value of "z1" should be roughly equal to "ze1" plus + "z2". + +o "m1" is the number of calls to rcu_try_flip_waitmb(). + +o "me1" is the number of times that rcu_try_flip_waitmb() finds + that at least one CPU has not yet executed a memory barrier. + +o "m2" is the number of times that rcu_try_flip_waitmb() finds that + all CPUs have executed a memory barrier. + + +Hierarchical RCU debugfs Files and Formats + +This implementation of RCU provides three debugfs files under the +top-level directory RCU: rcu/rcudata (which displays fields in struct +rcu_data), rcu/rcugp (which displays grace-period counters), and +rcu/rcuhier (which displays the struct rcu_node hierarchy). + +The output of "cat rcu/rcudata" looks as follows: + +rcu: + 0 c=1985 g=1986 pq=1 pqc=1985 qp=0 dt=26097 dn=2 df=9102 of=0 ri=11 ql=2 b=10 + 1 c=1985 g=1986 pq=1 pqc=1985 qp=0 dt=30421 dn=2 df=6608 of=0 ri=2 ql=39 b=10 + 2 c=1982 g=1982 pq=1 pqc=1982 qp=0 dt=10934 dn=2 df=9612 of=0 ri=0 ql=0 b=10 + 3 c=1985 g=1986 pq=1 pqc=1985 qp=0 dt=30139 dn=2 df=6043 of=0 ri=0 ql=58 b=10 + 4 c=1960 g=1960 pq=1 pqc=1960 qp=1 dt=1202 dn=2 df=30470 of=0 ri=3 ql=0 b=10 + 5 c=1985 g=1986 pq=1 pqc=1985 qp=0 dt=15341 dn=2 df=5350 of=0 ri=0 ql=25 b=10 + 6 c=1983 g=1984 pq=1 pqc=1983 qp=1 dt=516 dn=2 df=31950 of=0 ri=0 ql=0 b=10 + 7 c=1985 g=1986 pq=1 pqc=1985 qp=0 dt=8205 dn=2 df=7465 of=0 ri=0 ql=28 b=10 +rcu_bh: + 0 c=375 g=375 pq=1 pqc=375 qp=0 dt=26097 dn=2 df=0 of=0 ri=0 ql=0 b=10 + 1 c=375 g=375 pq=1 pqc=375 qp=0 dt=30421 dn=2 df=162 of=0 ri=0 ql=0 b=10 + 2 c=375 g=375 pq=1 pqc=375 qp=1 dt=10934 dn=2 df=162 of=0 ri=0 ql=0 b=10 + 3 c=375 g=375 pq=1 pqc=375 qp=0 dt=30139 dn=2 df=107 of=0 ri=0 ql=0 b=10 + 4 c=375 g=375 pq=1 pqc=375 qp=1 dt=1202 dn=2 df=174 of=0 ri=0 ql=0 b=10 + 5 c=375 g=375 pq=1 pqc=375 qp=0 dt=15341 dn=2 df=122 of=0 ri=0 ql=0 b=10 + 6 c=375 g=375 pq=1 pqc=375 qp=1 dt=516 dn=2 df=117 of=0 ri=0 ql=0 b=10 + 7 c=375 g=375 pq=1 pqc=375 qp=0 dt=8205 dn=2 df=127 of=0 ri=0 ql=0 b=10 + +The first section lists the rcu_data structures for rcu, the second for +rcu_bh. Each section has one line per CPU, or eight for this 8-CPU system. +The fields are as follows: + +o The number at the beginning of each line is the CPU number. + CPUs numbers followed by an exclamation mark are offline, + but have been online at least once since boot. There will be + no output for CPUs that have never been online, which can be + a good thing in the surprisingly common case where NR_CPUS is + substantially larger than the number of actual CPUs. + +o "c" is the count of grace periods that this CPU believes have + completed. CPUs in dynticks idle mode may lag quite a ways + behind, for example, CPU 4 under "rcu" above, which has slept + through the past 25 RCU grace periods. It is not unusual to + see CPUs lagging by thousands of grace periods. + +o "g" is the count of grace periods that this CPU believes have + started. Again, CPUs in dynticks idle mode may lag behind. + If the "c" and "g" values are equal, this CPU has already + reported a quiescent state for the last RCU grace period that + it is aware of, otherwise, the CPU believes that it owes RCU a + quiescent state. + +o "pq" indicates that this CPU has passed through a quiescent state + for the current grace period. It is possible for "pq" to be + "1" and "c" different than "g", which indicates that although + the CPU has passed through a quiescent state, either (1) this + CPU has not yet reported that fact, (2) some other CPU has not + yet reported for this grace period, or (3) both. + +o "pqc" indicates which grace period the last-observed quiescent + state for this CPU corresponds to. This is important for handling + the race between CPU 0 reporting an extended dynticks-idle + quiescent state for CPU 1 and CPU 1 suddenly waking up and + reporting its own quiescent state. If CPU 1 was the last CPU + for the current grace period, then the CPU that loses this race + will attempt to incorrectly mark CPU 1 as having checked in for + the next grace period! + +o "qp" indicates that RCU still expects a quiescent state from + this CPU. + +o "dt" is the current value of the dyntick counter that is incremented + when entering or leaving dynticks idle state, either by the + scheduler or by irq. + + This field is displayed only for CONFIG_NO_HZ kernels. + +o "dn" is the current value of the dyntick counter that is incremented + when entering or leaving dynticks idle state via NMI. If both + the "dt" and "dn" values are even, then this CPU is in dynticks + idle mode and may be ignored by RCU. If either of these two + counters is odd, then RCU must be alert to the possibility of + an RCU read-side critical section running on this CPU. + + This field is displayed only for CONFIG_NO_HZ kernels. + +o "df" is the number of times that some other CPU has forced a + quiescent state on behalf of this CPU due to this CPU being in + dynticks-idle state. + + This field is displayed only for CONFIG_NO_HZ kernels. + +o "of" is the number of times that some other CPU has forced a + quiescent state on behalf of this CPU due to this CPU being + offline. In a perfect world, this might neve happen, but it + turns out that offlining and onlining a CPU can take several grace + periods, and so there is likely to be an extended period of time + when RCU believes that the CPU is online when it really is not. + Please note that erring in the other direction (RCU believing a + CPU is offline when it is really alive and kicking) is a fatal + error, so it makes sense to err conservatively. + +o "ri" is the number of times that RCU has seen fit to send a + reschedule IPI to this CPU in order to get it to report a + quiescent state. + +o "ql" is the number of RCU callbacks currently residing on + this CPU. This is the total number of callbacks, regardless + of what state they are in (new, waiting for grace period to + start, waiting for grace period to end, ready to invoke). + +o "b" is the batch limit for this CPU. If more than this number + of RCU callbacks is ready to invoke, then the remainder will + be deferred. + + +The output of "cat rcu/rcudata" looks as follows: + +rcu: completed=33062 gpnum=33063 +rcu_bh: completed=464 gpnum=464 + +Again, this output is for both "rcu" and "rcu_bh". The fields are +taken from the rcu_state structure, and are as follows: + +o "completed" is the number of grace periods that have completed. + It is comparable to the "c" field from rcu/rcudata in that a + CPU whose "c" field matches the value of "completed" is aware + that the corresponding RCU grace period has completed. + +o "gpnum" is the number of grace periods that have started. It is + comparable to the "g" field from rcu/rcudata in that a CPU + whose "g" field matches the value of "gpnum" is aware that the + corresponding RCU grace period has started. + + If these two fields are equal (as they are for "rcu_bh" above), + then there is no grace period in progress, in other words, RCU + is idle. On the other hand, if the two fields differ (as they + do for "rcu" above), then an RCU grace period is in progress. + + +The output of "cat rcu/rcuhier" looks as follows, with very long lines: + +rcu: +c=33184 g=33185 s=0 jfq=1 nfqs=61601/nfqsng=28011(33590) +1/1 0:127 ^0 +1/3 0:35 ^0 0/0 36:71 ^1 0/0 72:107 ^2 0/0 108:127 ^3 +14/3f 0:5 ^0 0/3 6:11 ^1 0/0 12:17 ^2 0/0 18:23 ^3 0/0 24:29 ^4 0/0 30:35 ^5 0/0 36:41 ^0 0/0 42:47 ^1 0/0 48:53 ^2 0/0 54:59 ^3 0/0 60:65 ^4 0/0 66:71 ^5 0/0 72:77 ^0 0/0 78:83 ^1 0/0 84:89 ^2 0/0 90:95 ^3 0/0 96:101 ^4 0/0 102:107 ^5 0/0 108:113 ^0 0/0 114:119 ^1 0/0 120:125 ^2 0/0 126:127 ^3 +rcu_bh: +c=470 g=470 s=0 jfq=2 nfqs=62302/nfqsng=62027(275) +0/1 0:127 ^0 +0/3 0:35 ^0 0/0 36:71 ^1 0/0 72:107 ^2 0/0 108:127 ^3 +0/3f 0:5 ^0 0/3 6:11 ^1 0/0 12:17 ^2 0/0 18:23 ^3 0/0 24:29 ^4 0/0 30:35 ^5 0/0 36:41 ^0 0/0 42:47 ^1 0/0 48:53 ^2 0/0 54:59 ^3 0/0 60:65 ^4 0/0 66:71 ^5 0/0 72:77 ^0 0/0 78:83 ^1 0/0 84:89 ^2 0/0 90:95 ^3 0/0 96:101 ^4 0/0 102:107 ^5 0/0 108:113 ^0 0/0 114:119 ^1 0/0 120:125 ^2 0/0 126:127 ^3 + +This is once again split into "rcu" and "rcu_bh" portions. The fields are +as follows: + +o "c" is exactly the same as "completed" under rcu/rcugp. + +o "g" is exactly the same as "gpnum" under rcu/rcugp. + +o "s" is the "signaled" state that drives force_quiescent_state()'s + state machine. + +o "jfq" is the number of jiffies remaining for this grace period + before force_quiescent_state() is invoked to help push things + along. Note that CPUs in dyntick-idle mode thoughout the grace + period will not report on their own, but rather must be check by + some other CPU via force_quiescent_state(). + +o "nfqs" is the number of calls to force_quiescent_state() since + boot. + +o "nfqsng" is the number of useless calls to force_quiescent_state(), + where there wasn't actually a grace period active. This can + happen due to races. The number in parentheses is the difference + between "nfqs" and "nfqsng", or the number of times that + force_quiescent_state() actually did some real work. + +o Each element of the form "1/1 0:127 ^0" represents one struct + rcu_node. Each line represents one level of the hierarchy, from + root to leaves. It is best to think of the rcu_data structures + as forming yet another level after the leaves. Note that there + might be either one, two, or three levels of rcu_node structures, + depending on the relationship between CONFIG_RCU_FANOUT and + CONFIG_NR_CPUS. + + o The numbers separated by the "/" are the qsmask followed + by the qsmaskinit. The qsmask will have one bit + set for each entity in the next lower level that + has not yet checked in for the current grace period. + The qsmaskinit will have one bit for each entity that is + currently expected to check in during each grace period. + The value of qsmaskinit is assigned to that of qsmask + at the beginning of each grace period. + + For example, for "rcu", the qsmask of the first entry + of the lowest level is 0x14, meaning that we are still + waiting for CPUs 2 and 4 to check in for the current + grace period. + + o The numbers separated by the ":" are the range of CPUs + served by this struct rcu_node. This can be helpful + in working out how the hierarchy is wired together. + + For example, the first entry at the lowest level shows + "0:5", indicating that it covers CPUs 0 through 5. + + o The number after the "^" indicates the bit in the + next higher level rcu_node structure that this + rcu_node structure corresponds to. + + For example, the first entry at the lowest level shows + "^0", indicating that it corresponds to bit zero in + the first entry at the middle level. diff --git a/arch/powerpc/platforms/pseries/rtasd.c b/arch/powerpc/platforms/pseries/rtasd.c index c9ffd8c..d8e784a 100644 --- a/arch/powerpc/platforms/pseries/rtasd.c +++ b/arch/powerpc/platforms/pseries/rtasd.c @@ -208,6 +208,7 @@ void pSeries_log_error(char *buf, unsigned int err_type, int fatal) break; case ERR_TYPE_KERNEL_PANIC: default: + WARN_ON_ONCE(!irqs_disabled()); /* @@@ DEBUG @@@ */ spin_unlock_irqrestore(&rtasd_log_lock, s); return; } @@ -227,6 +228,7 @@ void pSeries_log_error(char *buf, unsigned int err_type, int fatal) /* Check to see if we need to or have stopped logging */ if (fatal || !logging_enabled) { logging_enabled = 0; + WARN_ON_ONCE(!irqs_disabled()); /* @@@ DEBUG @@@ */ spin_unlock_irqrestore(&rtasd_log_lock, s); return; } @@ -249,11 +251,13 @@ void pSeries_log_error(char *buf, unsigned int err_type, int fatal) else rtas_log_start += 1; + WARN_ON_ONCE(!irqs_disabled()); /* @@@ DEBUG @@@ */ spin_unlock_irqrestore(&rtasd_log_lock, s); wake_up_interruptible(&rtas_log_wait); break; case ERR_TYPE_KERNEL_PANIC: default: + WARN_ON_ONCE(!irqs_disabled()); /* @@@ DEBUG @@@ */ spin_unlock_irqrestore(&rtasd_log_lock, s); return; } diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index 181006c..9b70b92 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -118,13 +118,17 @@ static inline void account_system_vtime(struct task_struct *tsk) } #endif -#if defined(CONFIG_PREEMPT_RCU) && defined(CONFIG_NO_HZ) +#if defined(CONFIG_NO_HZ) && !defined(CONFIG_CLASSIC_RCU) extern void rcu_irq_enter(void); extern void rcu_irq_exit(void); +extern void rcu_nmi_enter(void); +extern void rcu_nmi_exit(void); #else # define rcu_irq_enter() do { } while (0) # define rcu_irq_exit() do { } while (0) -#endif /* CONFIG_PREEMPT_RCU */ +# define rcu_nmi_enter() do { } while (0) +# define rcu_nmi_exit() do { } while (0) +#endif /* #if defined(CONFIG_NO_HZ) && !defined(CONFIG_CLASSIC_RCU) */ /* * It is safe to do non-atomic ops on ->hardirq_context, @@ -134,7 +138,6 @@ extern void rcu_irq_exit(void); */ #define __irq_enter() \ do { \ - rcu_irq_enter(); \ account_system_vtime(current); \ add_preempt_count(HARDIRQ_OFFSET); \ trace_hardirq_enter(); \ @@ -153,7 +156,6 @@ extern void irq_enter(void); trace_hardirq_exit(); \ account_system_vtime(current); \ sub_preempt_count(HARDIRQ_OFFSET); \ - rcu_irq_exit(); \ } while (0) /* @@ -161,7 +163,7 @@ extern void irq_enter(void); */ extern void irq_exit(void); -#define nmi_enter() do { lockdep_off(); __irq_enter(); } while (0) -#define nmi_exit() do { __irq_exit(); lockdep_on(); } while (0) +#define nmi_enter() do { lockdep_off(); rcu_nmi_enter(); __irq_enter(); } while (0) +#define nmi_exit() do { __irq_exit(); rcu_nmi_exit(); lockdep_on(); } while (0) #endif /* LINUX_HARDIRQ_H */ diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h index e8b4039..f8544ae 100644 --- a/include/linux/rcupdate.h +++ b/include/linux/rcupdate.h @@ -52,11 +52,15 @@ struct rcu_head { void (*func)(struct rcu_head *head); }; -#ifdef CONFIG_CLASSIC_RCU +#if defined(CONFIG_CLASSIC_RCU) #include <linux/rcuclassic.h> -#else /* #ifdef CONFIG_CLASSIC_RCU */ +#elif defined(CONFIG_TREE_RCU) +#include <linux/rcutree.h> +#elif defined(CONFIG_PREEMPT_RCU) #include <linux/rcupreempt.h> -#endif /* #else #ifdef CONFIG_CLASSIC_RCU */ +#else +#error "Unknown RCU implementation specified to kernel configuration" +#endif /* #else #if defined(CONFIG_CLASSIC_RCU) */ #define RCU_HEAD_INIT { .next = NULL, .func = NULL } #define RCU_HEAD(head) struct rcu_head head = RCU_HEAD_INIT diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h new file mode 100644 index 0000000..7f8e70b --- /dev/null +++ b/include/linux/rcutree.h @@ -0,0 +1,320 @@ +/* + * 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 */ + + /* 3) rcu-barrier functions */ + struct rcu_head barrier; + +#ifdef CONFIG_NO_HZ + /* 4) dynticks interface (see http://lwn.net/Articles/279077/) */ + 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 */ + + /* 5) 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. */ + + int cpu; +}; + +/* Values for signaled field in struc rcu_data. */ +#define RCU_SAVE_DYNTICK 0 /* Need to scan dyntick state. */ +#define RCU_FORCE_QS 1 /* 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 3 /* for rsp->seconds_stall */ +#define RCU_SECONDS_TILL_STALL_RECHECK 30 /* for rsp->seconds_stall */ +#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_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 seconds_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, _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 */ diff --git a/init/Kconfig b/init/Kconfig index b678803..6fdca78 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -914,10 +914,16 @@ source "block/Kconfig" config PREEMPT_NOTIFIERS bool -config CLASSIC_RCU - def_bool !PREEMPT_RCU +config TREE_RCU_TRACE + def_bool RCU_TRACE && TREE_RCU + select DEBUG_FS help - This option selects the classic RCU implementation that is - designed for best read-side performance on non-realtime - systems. Classic RCU is the default. Note that the - PREEMPT_RCU symbol is used to select/deselect this option. + This option provides tracing for the TREE_RCU implementation, + permitting Makefile to trivially select kernel/rcutree_trace.c. + +config PREEMPT_RCU_TRACE + def_bool RCU_TRACE && PREEMPT_RCU + select DEBUG_FS + help + This option provides tracing for the PREEMPT_RCU implementation, + permitting Makefile to trivially select kernel/rcupreempt_trace.c. diff --git a/kernel/Kconfig.preempt b/kernel/Kconfig.preempt index 9fdba03..463f297 100644 --- a/kernel/Kconfig.preempt +++ b/kernel/Kconfig.preempt @@ -52,10 +52,29 @@ config PREEMPT endchoice +choice + prompt "RCU Implementation" + default CLASSIC_RCU + +config CLASSIC_RCU + bool "Classic RCU" + help + This option selects the classic RCU implementation that is + designed for best read-side performance on non-realtime + systems. + + Select this option if you are unsure. + +config TREE_RCU + bool "Tree-based hierarchical RCU" + help + This option selects the RCU implementation that is + designed for very large SMP system with hundreds or + thousands of CPUs. + config PREEMPT_RCU bool "Preemptible RCU" depends on PREEMPT - default n help This option reduces the latency of the kernel by making certain RCU sections preemptible. Normally RCU code is non-preemptible, if @@ -64,16 +83,47 @@ config PREEMPT_RCU now-naive assumptions about each RCU read-side critical section remaining on a given CPU through its execution. - Say N if you are unsure. +endchoice config RCU_TRACE - bool "Enable tracing for RCU - currently stats in debugfs" - depends on PREEMPT_RCU - select DEBUG_FS - default y + bool "Enable tracing for RCU" + depends on TREE_RCU || PREEMPT_RCU help This option provides tracing in RCU which presents stats in debugfs for debugging RCU implementation. Say Y here if you want to enable RCU tracing Say N if you are unsure. + +config RCU_FANOUT + int "Tree-based hierarchical RCU fanout value" + range 2 64 if 64BIT + range 2 32 if !64BIT + depends on TREE_RCU + default 64 if 64BIT + default 32 if !64BIT + help + This option controls the fanout of hierarchical implementations + of RCU, allowing RCU to work efficiently on machines with + large numbers of CPUs. This value must be at least the cube + root of NR_CPUS, which allows NR_CPUS up to 32,768 for 32-bit + systems and up to 262,144 for 64-bit systems. + + Select a specific number if testing RCU itself. + Take the default if unsure. + +config RCU_FANOUT_EXACT + bool "Disable tree-based hierarchical RCU auto-balancing" + depends on TREE_RCU + default n + help + This option forces use of the exact RCU_FANOUT value specified, + regardless of imbalances in the hierarchy. This is useful for + testing RCU itself, and might one day be useful on systems with + strong NUMA behavior. + + Without RCU_FANOUT_EXACT, the code will balance the hierarchy. + + Say n if unsure. + + diff --git a/kernel/Makefile b/kernel/Makefile index 4e1d7df..101e880 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -74,10 +74,10 @@ obj-$(CONFIG_GENERIC_HARDIRQS) += irq/ obj-$(CONFIG_SECCOMP) += seccomp.o obj-$(CONFIG_RCU_TORTURE_TEST) += rcutorture.o obj-$(CONFIG_CLASSIC_RCU) += rcuclassic.o +obj-$(CONFIG_TREE_RCU) += rcutree.o obj-$(CONFIG_PREEMPT_RCU) += rcupreempt.o -ifeq ($(CONFIG_PREEMPT_RCU),y) -obj-$(CONFIG_RCU_TRACE) += rcupreempt_trace.o -endif +obj-$(CONFIG_TREE_RCU_TRACE) += rcutree_trace.o +obj-$(CONFIG_PREEMPT_RCU_TRACE) += rcupreempt_trace.o obj-$(CONFIG_RELAY) += relay.o obj-$(CONFIG_SYSCTL) += utsname_sysctl.o obj-$(CONFIG_TASK_DELAY_ACCT) += delayacct.o diff --git a/kernel/rcupreempt.c b/kernel/rcupreempt.c index 2782793..6bc8489 100644 --- a/kernel/rcupreempt.c +++ b/kernel/rcupreempt.c @@ -559,6 +559,16 @@ void rcu_irq_exit(void) } } +void rcu_nmi_enter(void) +{ + rcu_irq_enter(); +} + +void rcu_nmi_exit(void) +{ + rcu_irq_exit(); +} + static void dyntick_save_progress_counter(int cpu) { struct rcu_dyntick_sched *rdssp = &per_cpu(rcu_dyntick_sched, cpu); diff --git a/kernel/rcupreempt_trace.c b/kernel/rcupreempt_trace.c index 5edf82c..def42e8 100644 --- a/kernel/rcupreempt_trace.c +++ b/kernel/rcupreempt_trace.c @@ -149,12 +149,12 @@ static void rcupreempt_trace_sum(struct rcupreempt_trace *sp) sp->done_length += cp->done_length; sp->done_add += cp->done_add; sp->done_remove += cp->done_remove; - atomic_set(&sp->done_invoked, atomic_read(&cp->done_invoked)); + atomic_add(atomic_read(&cp->done_invoked), &sp->done_invoked); sp->rcu_check_callbacks += cp->rcu_check_callbacks; - atomic_set(&sp->rcu_try_flip_1, - atomic_read(&cp->rcu_try_flip_1)); - atomic_set(&sp->rcu_try_flip_e1, - atomic_read(&cp->rcu_try_flip_e1)); + atomic_add(atomic_read(&cp->rcu_try_flip_1), + &sp->rcu_try_flip_1); + atomic_add(atomic_read(&cp->rcu_try_flip_e1), + &sp->rcu_try_flip_e1); sp->rcu_try_flip_i1 += cp->rcu_try_flip_i1; sp->rcu_try_flip_ie1 += cp->rcu_try_flip_ie1; sp->rcu_try_flip_g1 += cp->rcu_try_flip_g1; diff --git a/kernel/rcutree.c b/kernel/rcutree.c new file mode 100644 index 0000000..ed4e1f0 --- /dev/null +++ b/kernel/rcutree.c @@ -0,0 +1,1502 @@ +/* + * Read-Copy Update mechanism for mutual exclusion + * + * 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 + * + * Authors: Dipankar Sarma <dipankar@in.ibm.com> + * Manfred Spraul <manfred@colorfullife.com> + * Paul E. McKenney <paulmck@linux.vnet.ibm.com> Hierarchical version + * + * 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 + */ +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/spinlock.h> +#include <linux/smp.h> +#include <linux/rcupdate.h> +#include <linux/interrupt.h> +#include <linux/sched.h> +#include <asm/atomic.h> +#include <linux/bitops.h> +#include <linux/module.h> +#include <linux/completion.h> +#include <linux/moduleparam.h> +#include <linux/percpu.h> +#include <linux/notifier.h> +#include <linux/cpu.h> +#include <linux/mutex.h> +#include <linux/time.h> + +#ifdef CONFIG_DEBUG_LOCK_ALLOC +static struct lock_class_key rcu_lock_key; +struct lockdep_map rcu_lock_map = + STATIC_LOCKDEP_MAP_INIT("rcu_read_lock", &rcu_lock_key); +EXPORT_SYMBOL_GPL(rcu_lock_map); +#endif + +/* Data structures. */ + +#define RCU_STATE_INITIALIZER(name) { \ + .level = { &name.node[0] }, \ + .levelcnt = { \ + NUM_RCU_LVL_0, /* root of hierarchy. */ \ + NUM_RCU_LVL_1, \ + NUM_RCU_LVL_2, \ + NUM_RCU_LVL_3, /* == MAX_RCU_LVLS */ \ + }, \ + .signaled = RCU_SIGNAL_INIT, \ + .gpnum = -300, \ + .completed = -300, \ + .onofflock = __SPIN_LOCK_UNLOCKED(&name.onofflock), \ + .fqslock = __SPIN_LOCK_UNLOCKED(&name.fqslock), \ + .n_force_qs = 0, \ + .n_force_qs_ngp = 0, \ +} + +struct rcu_state rcu_state = RCU_STATE_INITIALIZER(rcu_state); +DEFINE_PER_CPU(struct rcu_data, rcu_data); + +struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh_state); +DEFINE_PER_CPU(struct rcu_data, rcu_bh_data); + +#ifdef CONFIG_NO_HZ +DEFINE_PER_CPU(struct rcu_dynticks, rcu_dynticks); +#endif /* #ifdef CONFIG_NO_HZ */ + +static int blimit = 10; /* Maximum callbacks per softirq. */ +static int qhimark = 10000; /* If this many pending, ignore blimit. */ +static int qlowmark = 100; /* Once only this many pending, use blimit. */ + +static void force_quiescent_state(struct rcu_state *rsp, int relaxed); + +/* + * Return the number of RCU batches processed thus far for debug & stats. + */ +long rcu_batches_completed(void) +{ + return rcu_state.completed; +} +EXPORT_SYMBOL_GPL(rcu_batches_completed); + +/* + * Return the number of RCU BH batches processed thus far for debug & stats. + */ +long rcu_batches_completed_bh(void) +{ + return rcu_bh_state.completed; +} +EXPORT_SYMBOL_GPL(rcu_batches_completed_bh); + +/* + * Does the CPU have callbacks ready to be invoked? + */ +static int +cpu_has_callbacks_ready_to_invoke(struct rcu_data *rdp) +{ + return &rdp->nxtlist != rdp->nxttail[RCU_DONE_TAIL]; +} + +/* + * Does the current CPU require a yet-as-unscheduled grace period? + */ +static int +cpu_needs_another_gp(struct rcu_state *rsp, struct rcu_data *rdp) +{ + /* ACCESS_ONCE() because we are accessing outside of lock. */ + return *rdp->nxttail[RCU_DONE_TAIL] && + ACCESS_ONCE(rsp->completed) == ACCESS_ONCE(rsp->gpnum); +} + +/* + * Return the root node of the specified rcu_state structure. + */ +static struct rcu_node *rcu_get_root(struct rcu_state *rsp) +{ + return &rsp->node[0]; +} + +/* + * If the specified CPU is offline, tell the caller that it is in + * a quiescent state. Otherwise, whack it with a reschedule IPI. + * Grace periods can end up waiting on an offline CPU when that + * CPU is in the process of coming online -- it will be added to the + * rcu_node bitmasks before it actually makes it online. Because this + * race is quite rare, we check for it after detecting that the grace + * period has been delayed rather than checking each and every CPU + * each and every time we start a new grace period. + */ +static int rcu_implicit_offline_qs(struct rcu_data *rdp) +{ + /* + * If the CPU is offline, it is in a quiescent state. We can + * trust its state not to change because interrupts are disabled. + */ + if (cpu_is_offline(rdp->cpu)) { + rdp->offline_fqs++; + return 1; + } + + /* The CPU is online, so send it a reschedule IPI. */ + if (rdp->cpu != smp_processor_id()) + smp_send_reschedule(rdp->cpu); + else + set_need_resched(); + rdp->resched_ipi++; + return 0; +} + +#ifdef CONFIG_NO_HZ +static DEFINE_RATELIMIT_STATE(rcu_rs, 10 * HZ, 5); + +/* + * Enter nohz mode, in other words, -leave- the mode in which RCU + * read-side critical sections can occur. (Though RCU read-side + * critical sections can occur in irq handlers in nohz mode, a possibility + * handled by rcu_irq_enter() and rcu_irq_exit()). + */ +void rcu_enter_nohz(void) +{ + unsigned long flags; + struct rcu_dynticks *rdtp; + + smp_mb(); /* CPUs seeing ++ must see prior RCU read-side crit sects */ + local_irq_save(flags); + rdtp = &__get_cpu_var(rcu_dynticks); + rdtp->dynticks++; + rdtp->dynticks_nesting++; + WARN_ON_RATELIMIT(__get_cpu_var(rcu_dynticks).dynticks & 0x1, &rcu_rs); + local_irq_restore(flags); +} + +/* + * Exit nohz mode. + */ +void rcu_exit_nohz(void) +{ + unsigned long flags; + struct rcu_dynticks *rdtp; + + local_irq_save(flags); + rdtp = &__get_cpu_var(rcu_dynticks); + rdtp->dynticks++; + rdtp->dynticks_nesting--; + WARN_ON_RATELIMIT(!(__get_cpu_var(rcu_dynticks).dynticks & 0x1), + &rcu_rs); + local_irq_restore(flags); + smp_mb(); /* CPUs seeing ++ must see later RCU read-side crit sects */ +} + +/** + * rcu_nmi_enter - Called from NMI + * + * If the CPU was idle with dynamic ticks active, and there is no + * irq handler running, this updates rdtp->dynticks_nmi to let the + * RCU grace-period handling know that the CPU is active. + */ +void rcu_nmi_enter(void) +{ + struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks); + + if (rdtp->dynticks & 0x1) + return; + rdtp->dynticks_nmi++; + WARN_ON_RATELIMIT(!(rdtp->dynticks_nmi & 0x1), &rcu_rs); +} + +/** + * rcu_nmi_exit - Called from NMI + * + * If the CPU was idle with dynamic ticks active, and there is no + * irq handler running, this updates rdtp->dynticks_nmi to let the + * RCU grace-period handling know that the CPU is no longer active. + */ +void rcu_nmi_exit(void) +{ + struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks); + + if (rdtp->dynticks & 0x1) + return; + rdtp->dynticks_nmi++; + WARN_ON_RATELIMIT(rdtp->dynticks_nmi & 0x1, &rcu_rs); +} + +/** + * rcu_irq_enter - Called from hard irq handlers + * + * If the CPU was idle with dynamic ticks active, this updates the + * rdtp->dynticks to let the RCU handling know that the CPU is active. + */ +void rcu_irq_enter(void) +{ + struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks); + + if (rdtp->dynticks_nesting++) { + return; + } + rdtp->dynticks++; + WARN_ON_RATELIMIT(!(rdtp->dynticks & 0x1), &rcu_rs); +} + +/** + * rcu_irq_exit - Called when exiting hard irq context. + * + * If the CPU was idle with dynamic ticks active, update the rdp->dynticks + * to put let the RCU handling be aware that the CPU is going back to idle + * with no ticks. + */ +void rcu_irq_exit(void) +{ + struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks); + + if (--rdtp->dynticks_nesting) { + return; + } + rdtp->dynticks++; + WARN_ON_RATELIMIT(rdtp->dynticks & 0x1, &rcu_rs); + + /* If the interrupt queued a callback, get out of dyntick mode. */ + if (__get_cpu_var(rcu_data).nxtlist || + __get_cpu_var(rcu_bh_data).nxtlist) + set_need_resched(); +} + +/* + * Record the specified "completed" value, which is later used to validate + * dynticks counter manipulations. Specify "rsp->complete - 1" to + * unconditionally invalidate any future dynticks manipulations (which is + * useful at the beginning of a grace period). + */ +static void dyntick_record_completed(struct rcu_state *rsp, int comp) +{ + rsp->dynticks_completed = comp; +} + +/* + * Recall the previously recorded value of the completion for dynticks. + */ +static long dyntick_recall_completed(struct rcu_state *rsp) +{ + return rsp->dynticks_completed; +} + +/* + * Snapshot the specified CPU's dynticks counter so that we can later + * credit them with an implicit quiescent state. Return 1 if this CPU + * is already in a quiescent state courtesy of dynticks idle mode. + */ +static int dyntick_save_progress_counter(struct rcu_data *rdp) +{ + int ret; + int snap; + int snap_nmi; + + snap = rdp->dynticks->dynticks; + snap_nmi = rdp->dynticks->dynticks_nmi; + smp_mb(); /* Order sampling of snap with end of grace period. */ + rdp->dynticks_snap = snap; + rdp->dynticks_nmi_snap = snap_nmi; + ret = ((snap & 0x1) == 0) && ((snap_nmi & 0x1) == 0); + if (ret) + rdp->dynticks_fqs++; + return ret; +} + +/* + * Return true if the specified CPU has passed through a quiescent + * state by virtue of being in or having passed through an dynticks + * idle state since the last call to dyntick_save_progress_counter() + * for this same CPU. + */ +static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) +{ + long curr; + long curr_nmi; + long snap; + long snap_nmi; + + curr = rdp->dynticks->dynticks; + snap = rdp->dynticks_snap; + curr_nmi = rdp->dynticks->dynticks_nmi; + snap_nmi = rdp->dynticks_nmi_snap; + smp_mb(); /* force ordering with cpu entering/leaving dynticks. */ + + /* + * If the CPU passed through or entered a dynticks idle phase with + * no active irq/NMI handlers, then we can safely pretend that the CPU + * already acknowledged the request to pass through a quiescent + * state. Either way, that CPU cannot possibly be in an RCU + * read-side critical section that started before the beginning + * of the current RCU grace period. + */ + if ((curr != snap || (curr & 0x1) == 0) && + (curr_nmi != snap_nmi || (curr_nmi & 0x1) == 0)) { + rdp->dynticks_fqs++; + return 1; + } + + /* Go check for the CPU being offline. */ + return rcu_implicit_offline_qs(rdp); +} + +#else /* #ifdef CONFIG_NO_HZ */ + +static void dyntick_record_completed(struct rcu_state *rsp, int comp) +{ +} + +/* + * If there are no dynticks, then the only way that a CPU can passively + * be in a quiescent state is to be offline. Unlike dynticks idle, which + * is a point in time during the prior (already finished) grace period, + * an offline CPU is always in a quiescent state, and thus can be + * unconditionally applied. So just return the current value of completed. + */ +static long dyntick_recall_completed(struct rcu_state *rsp) +{ + return rsp->completed; +} + +static int dyntick_save_progress_counter(struct rcu_data *rdp) { return 0; } + +static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) +{ + return rcu_implicit_offline_qs(rdp); +} + +#endif /* #else #ifdef CONFIG_NO_HZ */ + +#ifdef CONFIG_RCU_CPU_STALL_DETECTOR + +static void record_gp_stall_check_time(struct rcu_state *rsp) +{ + rsp->gp_start = jiffies; + rsp->seconds_stall = get_seconds() + RCU_SECONDS_TILL_STALL_CHECK; +} + +static void print_other_cpu_stall(struct rcu_state *rsp) +{ + int cpu; + long delta; + unsigned long flags; + struct rcu_node *rnp = rcu_get_root(rsp); + struct rcu_node *rnp_cur = rsp->level[NUM_RCU_LVLS - 1]; + struct rcu_node *rnp_end = &rsp->node[NUM_RCU_NODES]; + + /* Only let one CPU complain about others per time interval. */ + + spin_lock_irqsave(&rnp->lock, flags); + delta = get_seconds() - rsp->seconds_stall; + if (delta < 2 || rsp->gpnum != rsp->completed) { + spin_unlock_irqrestore(&rnp->lock, flags); + return; + } + rsp->seconds_stall = get_seconds() + RCU_SECONDS_TILL_STALL_RECHECK; + spin_unlock_irqrestore(&rnp->lock, flags); + + /* OK, time to rat on our buddy... */ + + printk(KERN_ERR "RCU detected CPU stalls:"); + for (; rnp_cur < rnp_end; rnp_cur++) { + if (rnp_cur->qsmask == 0) + continue; + for (cpu = 0; cpu <= rnp_cur->grphi - rnp_cur->grplo; cpu++) + if (rnp_cur->qsmask & (1UL << cpu)) + printk(" %d", rnp_cur->grplo + cpu); + } + printk(" (detected by %d, t=%ld jiffies)\n", + smp_processor_id(), (long)(jiffies - rsp->gp_start)); + force_quiescent_state(rsp, 0); /* Kick them all. */ +} + +static void print_cpu_stall(struct rcu_state *rsp) +{ + unsigned long flags; + struct rcu_node *rnp = rcu_get_root(rsp); + + printk(KERN_ERR "RCU detected CPU %d stall (t=%lu/%lu)\n", + smp_processor_id(), get_seconds(), + jiffies - rsp->gp_start); + dump_stack(); + spin_lock_irqsave(&rnp->lock, flags); + if ((long)(get_seconds() - rsp->seconds_stall) >= 0) + rsp->seconds_stall = + get_seconds() + RCU_SECONDS_TILL_STALL_RECHECK; + spin_unlock_irqrestore(&rnp->lock, flags); + set_need_resched(); /* kick ourselves to get things going. */ +} + +static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp) +{ + long delta; + struct rcu_node *rnp; + + delta = get_seconds() - rsp->seconds_stall; + rnp = rdp->mynode; + if ((rnp->qsmask & rdp->grpmask) && delta >= 0) { + + /* We haven't checked in, so go dump stack. */ + print_cpu_stall(rsp); + + } else if (rsp->gpnum != rsp->completed && delta >= 2) { + + /* They had two seconds to dump stack, so complain. */ + print_other_cpu_stall(rsp); + } +} + +#else /* #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ + +static void record_gp_stall_check_time(struct rcu_state *rsp) +{ +} + +static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp) +{ +} + +#endif /* #else #ifdef CONFIG_RCU_CPU_STALL_DETECTOR */ + +/* + * Update CPU-local rcu_data state to record the newly noticed grace period. + * This is used both when we started the grace period and when we notice + * that someone else started the grace period. + */ +static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) +{ + rdp->qs_pending = 1; + rdp->passed_quiesc = 0; + rdp->gpnum = rsp->gpnum; +} + +/* + * Did someone else start a new RCU grace period start since we last + * checked? Update local state appropriately if so. Must be called + * on the CPU corresponding to rdp. + */ +static int +check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp) +{ + unsigned long flags; + int ret = 0; + + local_irq_save(flags); + if (rdp->gpnum != rsp->gpnum) { + note_new_gpnum(rsp, rdp); + ret = 1; + } + local_irq_restore(flags); + return ret; +} + +/* + * Start a new RCU grace period if warranted, re-initializing the hierarchy + * in preparation for detecting the next grace period. The caller must hold + * the root node's ->lock, which is released before return. Hard irqs must + * be disabled. + */ +static void +rcu_start_gp(struct rcu_state *rsp, unsigned long iflg) + __releases(rsp->rda[smp_processor_id()]->lock) +{ + unsigned long flags = iflg; + struct rcu_data *rdp = rsp->rda[smp_processor_id()]; + struct rcu_node *rnp = rcu_get_root(rsp); + struct rcu_node *rnp_cur; + struct rcu_node *rnp_end; + + if (!cpu_needs_another_gp(rsp, rdp)) { + spin_unlock_irqrestore(&rnp->lock, flags); + return; + } + + /* Advance to a new grace period and initialize state. */ + rsp->gpnum++; + rsp->signaled = RCU_SIGNAL_INIT; + rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; + record_gp_stall_check_time(rsp); + dyntick_record_completed(rsp, rsp->completed - 1); + note_new_gpnum(rsp, rdp); + + /* + * Because we are first, we know that all our callbacks will + * be covered by this upcoming grace period, even the ones + * that were registered arbitrarily recently. + */ + rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL]; + rdp->nxttail[RCU_WAIT_TAIL] = rdp->nxttail[RCU_NEXT_TAIL]; + + /* Special-case the common single-level case. */ + if (NUM_RCU_NODES == 1) { + rnp->qsmask = rnp->qsmaskinit; + spin_unlock_irqrestore(&rnp->lock, flags); + return; + } + + spin_unlock_irqrestore(&rnp->lock, flags); + + + /* Exclude any concurrent CPU-hotplug operations. */ + spin_lock_irqsave(&rsp->onofflock, flags); + + /* + * Set the quiescent-state-needed bits in all the non-leaf RCU + * nodes for all currently online CPUs. This operation relies + * on the layout of the hierarchy within the rsp->node[] array. + * Note that other CPUs will access only the leaves of the + * hierarchy, which still indicate that no grace period is in + * progress. In addition, we have excluded CPU-hotplug operations. + * + * We therefore do not need to hold any locks. Any required + * memory barriers will be supplied by the locks guarding the + * leaf rcu_nodes in the hierarchy. + */ + + rnp_end = rsp->level[NUM_RCU_LVLS - 1]; + for (rnp_cur = &rsp->node[0]; rnp_cur < rnp_end; rnp_cur++) + rnp_cur->qsmask = rnp_cur->qsmaskinit; + + /* + * Now set up the leaf nodes. Here we must be careful. First, + * we need to hold the lock in order to exclude other CPUs, which + * might be contending for the leaf nodes' locks. Second, as + * soon as we initialize a given leaf node, its CPUs might run + * up the rest of the hierarchy. We must therefore acquire locks + * for each nod
