linux/kernel/smp.c

649 lines
17 KiB
C
Raw Normal View History

/*
* Generic helpers for smp ipi calls
*
* (C) Jens Axboe <jens.axboe@oracle.com> 2008
*/
#include <linux/rcupdate.h>
#include <linux/rculist.h>
#include <linux/kernel.h>
#include <linux/export.h>
#include <linux/percpu.h>
#include <linux/init.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
#include <linux/gfp.h>
#include <linux/smp.h>
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
#include <linux/cpu.h>
#include "smpboot.h"
enum {
CSD_FLAG_LOCK = 0x01,
CSD_FLAG_WAIT = 0x02,
};
struct call_function_data {
smp: make smp_call_function_many() use logic similar to smp_call_function_single() I'm testing swapout workload in a two-socket Xeon machine. The workload has 10 threads, each thread sequentially accesses separate memory region. TLB flush overhead is very big in the workload. For each page, page reclaim need move it from active lru list and then unmap it. Both need a TLB flush. And this is a multthread workload, TLB flush happens in 10 CPUs. In X86, TLB flush uses generic smp_call)function. So this workload stress smp_call_function_many heavily. Without patch, perf shows: + 24.49% [k] generic_smp_call_function_interrupt - 21.72% [k] _raw_spin_lock - _raw_spin_lock + 79.80% __page_check_address + 6.42% generic_smp_call_function_interrupt + 3.31% get_swap_page + 2.37% free_pcppages_bulk + 1.75% handle_pte_fault + 1.54% put_super + 1.41% grab_super_passive + 1.36% __swap_duplicate + 0.68% blk_flush_plug_list + 0.62% swap_info_get + 6.55% [k] flush_tlb_func + 6.46% [k] smp_call_function_many + 5.09% [k] call_function_interrupt + 4.75% [k] default_send_IPI_mask_sequence_phys + 2.18% [k] find_next_bit swapout throughput is around 1300M/s. With the patch, perf shows: - 27.23% [k] _raw_spin_lock - _raw_spin_lock + 80.53% __page_check_address + 8.39% generic_smp_call_function_single_interrupt + 2.44% get_swap_page + 1.76% free_pcppages_bulk + 1.40% handle_pte_fault + 1.15% __swap_duplicate + 1.05% put_super + 0.98% grab_super_passive + 0.86% blk_flush_plug_list + 0.57% swap_info_get + 8.25% [k] default_send_IPI_mask_sequence_phys + 7.55% [k] call_function_interrupt + 7.47% [k] smp_call_function_many + 7.25% [k] flush_tlb_func + 3.81% [k] _raw_spin_lock_irqsave + 3.78% [k] generic_smp_call_function_single_interrupt swapout throughput is around 1400M/s. So there is around a 7% improvement, and total cpu utilization doesn't change. Without the patch, cfd_data is shared by all CPUs. generic_smp_call_function_interrupt does read/write cfd_data several times which will create a lot of cache ping-pong. With the patch, the data becomes per-cpu. The ping-pong is avoided. And from the perf data, this doesn't make call_single_queue lock contend. Next step is to remove generic_smp_call_function_interrupt() from arch code. Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 00:43:03 +00:00
struct call_single_data __percpu *csd;
cpumask_var_t cpumask;
};
static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_function_data, cfd_data);
static DEFINE_PER_CPU_SHARED_ALIGNED(struct llist_head, call_single_queue);
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
static int
hotplug_cfd(struct notifier_block *nfb, unsigned long action, void *hcpu)
{
long cpu = (long)hcpu;
struct call_function_data *cfd = &per_cpu(cfd_data, cpu);
switch (action) {
case CPU_UP_PREPARE:
case CPU_UP_PREPARE_FROZEN:
if (!zalloc_cpumask_var_node(&cfd->cpumask, GFP_KERNEL,
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
cpu_to_node(cpu)))
return notifier_from_errno(-ENOMEM);
smp: make smp_call_function_many() use logic similar to smp_call_function_single() I'm testing swapout workload in a two-socket Xeon machine. The workload has 10 threads, each thread sequentially accesses separate memory region. TLB flush overhead is very big in the workload. For each page, page reclaim need move it from active lru list and then unmap it. Both need a TLB flush. And this is a multthread workload, TLB flush happens in 10 CPUs. In X86, TLB flush uses generic smp_call)function. So this workload stress smp_call_function_many heavily. Without patch, perf shows: + 24.49% [k] generic_smp_call_function_interrupt - 21.72% [k] _raw_spin_lock - _raw_spin_lock + 79.80% __page_check_address + 6.42% generic_smp_call_function_interrupt + 3.31% get_swap_page + 2.37% free_pcppages_bulk + 1.75% handle_pte_fault + 1.54% put_super + 1.41% grab_super_passive + 1.36% __swap_duplicate + 0.68% blk_flush_plug_list + 0.62% swap_info_get + 6.55% [k] flush_tlb_func + 6.46% [k] smp_call_function_many + 5.09% [k] call_function_interrupt + 4.75% [k] default_send_IPI_mask_sequence_phys + 2.18% [k] find_next_bit swapout throughput is around 1300M/s. With the patch, perf shows: - 27.23% [k] _raw_spin_lock - _raw_spin_lock + 80.53% __page_check_address + 8.39% generic_smp_call_function_single_interrupt + 2.44% get_swap_page + 1.76% free_pcppages_bulk + 1.40% handle_pte_fault + 1.15% __swap_duplicate + 1.05% put_super + 0.98% grab_super_passive + 0.86% blk_flush_plug_list + 0.57% swap_info_get + 8.25% [k] default_send_IPI_mask_sequence_phys + 7.55% [k] call_function_interrupt + 7.47% [k] smp_call_function_many + 7.25% [k] flush_tlb_func + 3.81% [k] _raw_spin_lock_irqsave + 3.78% [k] generic_smp_call_function_single_interrupt swapout throughput is around 1400M/s. So there is around a 7% improvement, and total cpu utilization doesn't change. Without the patch, cfd_data is shared by all CPUs. generic_smp_call_function_interrupt does read/write cfd_data several times which will create a lot of cache ping-pong. With the patch, the data becomes per-cpu. The ping-pong is avoided. And from the perf data, this doesn't make call_single_queue lock contend. Next step is to remove generic_smp_call_function_interrupt() from arch code. Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 00:43:03 +00:00
cfd->csd = alloc_percpu(struct call_single_data);
if (!cfd->csd) {
free_cpumask_var(cfd->cpumask);
return notifier_from_errno(-ENOMEM);
}
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
break;
#ifdef CONFIG_HOTPLUG_CPU
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
case CPU_UP_CANCELED:
case CPU_UP_CANCELED_FROZEN:
case CPU_DEAD:
case CPU_DEAD_FROZEN:
free_cpumask_var(cfd->cpumask);
smp: make smp_call_function_many() use logic similar to smp_call_function_single() I'm testing swapout workload in a two-socket Xeon machine. The workload has 10 threads, each thread sequentially accesses separate memory region. TLB flush overhead is very big in the workload. For each page, page reclaim need move it from active lru list and then unmap it. Both need a TLB flush. And this is a multthread workload, TLB flush happens in 10 CPUs. In X86, TLB flush uses generic smp_call)function. So this workload stress smp_call_function_many heavily. Without patch, perf shows: + 24.49% [k] generic_smp_call_function_interrupt - 21.72% [k] _raw_spin_lock - _raw_spin_lock + 79.80% __page_check_address + 6.42% generic_smp_call_function_interrupt + 3.31% get_swap_page + 2.37% free_pcppages_bulk + 1.75% handle_pte_fault + 1.54% put_super + 1.41% grab_super_passive + 1.36% __swap_duplicate + 0.68% blk_flush_plug_list + 0.62% swap_info_get + 6.55% [k] flush_tlb_func + 6.46% [k] smp_call_function_many + 5.09% [k] call_function_interrupt + 4.75% [k] default_send_IPI_mask_sequence_phys + 2.18% [k] find_next_bit swapout throughput is around 1300M/s. With the patch, perf shows: - 27.23% [k] _raw_spin_lock - _raw_spin_lock + 80.53% __page_check_address + 8.39% generic_smp_call_function_single_interrupt + 2.44% get_swap_page + 1.76% free_pcppages_bulk + 1.40% handle_pte_fault + 1.15% __swap_duplicate + 1.05% put_super + 0.98% grab_super_passive + 0.86% blk_flush_plug_list + 0.57% swap_info_get + 8.25% [k] default_send_IPI_mask_sequence_phys + 7.55% [k] call_function_interrupt + 7.47% [k] smp_call_function_many + 7.25% [k] flush_tlb_func + 3.81% [k] _raw_spin_lock_irqsave + 3.78% [k] generic_smp_call_function_single_interrupt swapout throughput is around 1400M/s. So there is around a 7% improvement, and total cpu utilization doesn't change. Without the patch, cfd_data is shared by all CPUs. generic_smp_call_function_interrupt does read/write cfd_data several times which will create a lot of cache ping-pong. With the patch, the data becomes per-cpu. The ping-pong is avoided. And from the perf data, this doesn't make call_single_queue lock contend. Next step is to remove generic_smp_call_function_interrupt() from arch code. Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 00:43:03 +00:00
free_percpu(cfd->csd);
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
break;
#endif
};
return NOTIFY_OK;
}
static struct notifier_block hotplug_cfd_notifier = {
.notifier_call = hotplug_cfd,
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
};
generic-ipi: Fix kexec boot crash by initializing call_single_queue before enabling interrupts There is a problem that kdump(2nd kernel) sometimes hangs up due to a pending IPI from 1st kernel. Kernel panic occurs because IPI comes before call_single_queue is initialized. To fix the crash, rename init_call_single_data() to call_function_init() and call it in start_kernel() so that call_single_queue can be initialized before enabling interrupts. The details of the crash are: (1) 2nd kernel boots up (2) A pending IPI from 1st kernel comes when irqs are first enabled in start_kernel(). (3) Kernel tries to handle the interrupt, but call_single_queue is not initialized yet at this point. As a result, in the generic_smp_call_function_single_interrupt(), NULL pointer dereference occurs when list_replace_init() tries to access &q->list.next. Therefore this patch changes the name of init_call_single_data() to call_function_init() and calls it before local_irq_enable() in start_kernel(). Signed-off-by: Takao Indoh <indou.takao@jp.fujitsu.com> Reviewed-by: WANG Cong <xiyou.wangcong@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Milton Miller <miltonm@bga.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: kexec@lists.infradead.org Link: http://lkml.kernel.org/r/D6CBEE2F420741indou.takao@jp.fujitsu.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
2011-03-29 16:35:04 +00:00
void __init call_function_init(void)
{
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
void *cpu = (void *)(long)smp_processor_id();
int i;
for_each_possible_cpu(i)
init_llist_head(&per_cpu(call_single_queue, i));
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
hotplug_cfd(&hotplug_cfd_notifier, CPU_UP_PREPARE, cpu);
register_cpu_notifier(&hotplug_cfd_notifier);
}
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
/*
* csd_lock/csd_unlock used to serialize access to per-cpu csd resources
*
* For non-synchronous ipi calls the csd can still be in use by the
* previous function call. For multi-cpu calls its even more interesting
* as we'll have to ensure no other cpu is observing our csd.
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
*/
static void csd_lock_wait(struct call_single_data *csd)
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
{
while (csd->flags & CSD_FLAG_LOCK)
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
cpu_relax();
}
static void csd_lock(struct call_single_data *csd)
{
csd_lock_wait(csd);
csd->flags |= CSD_FLAG_LOCK;
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
/*
* prevent CPU from reordering the above assignment
* to ->flags with any subsequent assignments to other
* fields of the specified call_single_data structure:
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
*/
smp_mb();
}
static void csd_unlock(struct call_single_data *csd)
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
{
WARN_ON((csd->flags & CSD_FLAG_WAIT) && !(csd->flags & CSD_FLAG_LOCK));
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
/*
* ensure we're all done before releasing data:
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
*/
smp_mb();
csd->flags &= ~CSD_FLAG_LOCK;
}
/*
* Insert a previously allocated call_single_data element
* for execution on the given CPU. data must already have
* ->func, ->info, and ->flags set.
*/
static void generic_exec_single(int cpu, struct call_single_data *csd, int wait)
{
if (wait)
csd->flags |= CSD_FLAG_WAIT;
/*
generic IPI: simplify barriers and locking Simplify the barriers in generic remote function call interrupt code. Firstly, just unconditionally take the lock and check the list in the generic_call_function_single_interrupt IPI handler. As we've just taken an IPI here, the chances are fairly high that there will be work on the list for us, so do the locking unconditionally. This removes the tricky lockless list_empty check and dubious barriers. The change looks bigger than it is because it is just removing an outer loop. Secondly, clarify architecture specific IPI locking rules. Generic code has no tools to impose any sane ordering on IPIs if they go outside normal cache coherency, ergo the arch code must make them appear to obey cache coherency as a "memory operation" to initiate an IPI, and a "memory operation" to receive one. This way at least they can be reasoned about in generic code, and smp_mb used to provide ordering. The combination of these two changes means that explict barriers can be taken out of queue handling for the single case -- shared data is explicitly locked, and ipi ordering must conform to that, so no barriers needed. An extra barrier is needed in the many handler, so as to ensure we load the list element after the IPI is received. Does any architecture actually *need* these barriers? For the initiator I could see it, but for the handler I would be surprised. So the other thing we could do for simplicity is just to require that, rather than just matching with cache coherency, we just require a full barrier before generating an IPI, and after receiving an IPI. In which case, the smp_mb()s can go away. But just for now, we'll be on the safe side and use the barriers (they're in the slow case anyway). Signed-off-by: Nick Piggin <npiggin@suse.de> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: linux-arch@vger.kernel.org Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Suresh Siddha <suresh.b.siddha@intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 05:22:45 +00:00
* The list addition should be visible before sending the IPI
* handler locks the list to pull the entry off it because of
* normal cache coherency rules implied by spinlocks.
*
* If IPIs can go out of order to the cache coherency protocol
* in an architecture, sufficient synchronisation should be added
* to arch code to make it appear to obey cache coherency WRT
* locking and barrier primitives. Generic code isn't really
* equipped to do the right thing...
*/
if (llist_add(&csd->llist, &per_cpu(call_single_queue, cpu)))
arch_send_call_function_single_ipi(cpu);
if (wait)
csd_lock_wait(csd);
}
/*
* Invoked by arch to handle an IPI for call function single. Must be
* called from the arch with interrupts disabled.
*/
void generic_smp_call_function_single_interrupt(void)
{
struct llist_node *entry, *next;
/*
* Shouldn't receive this interrupt on a cpu that is not yet online.
*/
WARN_ON_ONCE(!cpu_online(smp_processor_id()));
entry = llist_del_all(&__get_cpu_var(call_single_queue));
entry = llist_reverse_order(entry);
while (entry) {
struct call_single_data *csd;
next = entry->next;
csd = llist_entry(entry, struct call_single_data, llist);
csd->func(csd->info);
csd_unlock(csd);
entry = next;
}
}
static DEFINE_PER_CPU_SHARED_ALIGNED(struct call_single_data, csd_data);
generic-ipi: use per cpu data for single cpu ipi calls The smp_call_function can be passed a wait parameter telling it to wait for all the functions running on other CPUs to complete before returning, or to return without waiting. Unfortunately, this is currently just a suggestion and not manditory. That is, the smp_call_function can decide not to return and wait instead. The reason for this is because it uses kmalloc to allocate storage to send to the called CPU and that CPU will free it when it is done. But if we fail to allocate the storage, the stack is used instead. This means we must wait for the called CPU to finish before continuing. Unfortunatly, some callers do no abide by this hint and act as if the non-wait option is mandatory. The MTRR code for instance will deadlock if the smp_call_function is set to wait. This is because the smp_call_function will wait for the other CPUs to finish their called functions, but those functions are waiting on the caller to continue. This patch changes the generic smp_call_function code to use per cpu variables if the allocation of the data fails for a single CPU call. The smp_call_function_many will fall back to the smp_call_function_single if it fails its alloc. The smp_call_function_single is modified to not force the wait state. Since we now are using a single data per cpu we must synchronize the callers to prevent a second caller modifying the data before the first called IPI functions complete. To do so, I added a flag to the call_single_data called CSD_FLAG_LOCK. When the single CPU is called (which can be called when a many call fails an alloc), we set the LOCK bit on this per cpu data. When the caller finishes it clears the LOCK bit. The caller must wait till the LOCK bit is cleared before setting it. When it is cleared, there is no IPI function using it. Signed-off-by: Steven Rostedt <srostedt@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jens Axboe <jens.axboe@oracle.com> Acked-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-01-29 15:08:01 +00:00
/*
* smp_call_function_single - Run a function on a specific CPU
* @func: The function to run. This must be fast and non-blocking.
* @info: An arbitrary pointer to pass to the function.
* @wait: If true, wait until function has completed on other CPUs.
*
* Returns 0 on success, else a negative status code.
*/
int smp_call_function_single(int cpu, smp_call_func_t func, void *info,
int wait)
{
generic-ipi: remove kmalloc() Remove the use of kmalloc() from the smp_call_function_*() calls. Steven's generic-ipi patch (d7240b98: generic-ipi: use per cpu data for single cpu ipi calls) started the discussion on the use of kmalloc() in this code and fixed the smp_call_function_single(.wait=0) fallback case. In this patch we complete this by also providing means for the _many() call, which fully removes the need for kmalloc() in this code. The problem with the _many() call is that other cpus might still be observing our entry when we're done with it. It solved this by dynamically allocating data elements and RCU-freeing it. We solve it by using a single per-cpu entry which provides static storage and solves one half of the problem (avoiding referencing freed data). The other half, ensuring the queue iteration it still possible, is done by placing re-used entries at the head of the list. This means that if someone was still iterating that entry when it got moved, he will now re-visit the entries on the list he had already seen, but avoids skipping over entries like would have happened had we placed the new entry at the end. Furthermore, visiting entries twice is not a problem, since we remove our cpu from the entry's cpumask once its called. Many thanks to Oleg for his suggestions and him poking holes in my earlier attempts. Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Nick Piggin <npiggin@suse.de> Cc: Jens Axboe <jens.axboe@oracle.com> Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-25 12:59:47 +00:00
struct call_single_data d = {
.flags = 0,
};
unsigned long flags;
int this_cpu;
int err = 0;
/*
* prevent preemption and reschedule on another processor,
* as well as CPU removal
*/
this_cpu = get_cpu();
/*
* Can deadlock when called with interrupts disabled.
* We allow cpu's that are not yet online though, as no one else can
* send smp call function interrupt to this cpu and as such deadlocks
* can't happen.
*/
WARN_ON_ONCE(cpu_online(this_cpu) && irqs_disabled()
&& !oops_in_progress);
if (cpu == this_cpu) {
local_irq_save(flags);
func(info);
local_irq_restore(flags);
} else {
if ((unsigned)cpu < nr_cpu_ids && cpu_online(cpu)) {
struct call_single_data *csd = &d;
if (!wait)
csd = &__get_cpu_var(csd_data);
csd_lock(csd);
csd->func = func;
csd->info = info;
generic_exec_single(cpu, csd, wait);
} else {
err = -ENXIO; /* CPU not online */
}
}
put_cpu();
return err;
}
EXPORT_SYMBOL(smp_call_function_single);
/*
* smp_call_function_any - Run a function on any of the given cpus
* @mask: The mask of cpus it can run on.
* @func: The function to run. This must be fast and non-blocking.
* @info: An arbitrary pointer to pass to the function.
* @wait: If true, wait until function has completed.
*
* Returns 0 on success, else a negative status code (if no cpus were online).
*
* Selection preference:
* 1) current cpu if in @mask
* 2) any cpu of current node if in @mask
* 3) any other online cpu in @mask
*/
int smp_call_function_any(const struct cpumask *mask,
smp_call_func_t func, void *info, int wait)
{
unsigned int cpu;
const struct cpumask *nodemask;
int ret;
/* Try for same CPU (cheapest) */
cpu = get_cpu();
if (cpumask_test_cpu(cpu, mask))
goto call;
/* Try for same node. */
nodemask = cpumask_of_node(cpu_to_node(cpu));
for (cpu = cpumask_first_and(nodemask, mask); cpu < nr_cpu_ids;
cpu = cpumask_next_and(cpu, nodemask, mask)) {
if (cpu_online(cpu))
goto call;
}
/* Any online will do: smp_call_function_single handles nr_cpu_ids. */
cpu = cpumask_any_and(mask, cpu_online_mask);
call:
ret = smp_call_function_single(cpu, func, info, wait);
put_cpu();
return ret;
}
EXPORT_SYMBOL_GPL(smp_call_function_any);
/**
generic-ipi: Fix deadlock in __smp_call_function_single Just got my 6 way machine to a state where cpu 0 is in an endless loop within __smp_call_function_single. All other cpus are idle. The call trace on cpu 0 looks like this: __smp_call_function_single scheduler_tick update_process_times tick_sched_timer __run_hrtimer hrtimer_interrupt clock_comparator_work do_extint ext_int_handler ----> timer irq cpu_idle __smp_call_function_single() got called from nohz_balancer_kick() (inlined) with the remote cpu being 1, wait being 0 and the per cpu variable remote_sched_softirq_cb (call_single_data) of the current cpu (0). Then it loops forever when it tries to grab the lock of the call_single_data, since it is already locked and enqueued on cpu 0. My theory how this could have happened: for some reason the scheduler decided to call __smp_call_function_single() on it's own cpu, and sends an IPI to itself. The interrupt stays pending since IRQs are disabled. If then the hypervisor schedules the cpu away it might happen that upon rescheduling both the IPI and the timer IRQ are pending. If then interrupts are enabled again it depends which one gets scheduled first. If the timer interrupt gets delivered first we end up with the local deadlock as seen in the calltrace above. Let's make __smp_call_function_single() check if the target cpu is the current cpu and execute the function immediately just like smp_call_function_single does. That should prevent at least the scenario described here. It might also be that the scheduler is not supposed to call __smp_call_function_single with the remote cpu being the current cpu, but that is a different issue. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jens Axboe <jaxboe@fusionio.com> Cc: Venkatesh Pallipadi <venki@google.com> Cc: Suresh Siddha <suresh.b.siddha@intel.com> LKML-Reference: <20100910114729.GB2827@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-10 11:47:29 +00:00
* __smp_call_function_single(): Run a function on a specific CPU
* @cpu: The CPU to run on.
* @data: Pre-allocated and setup data structure
generic-ipi: Fix deadlock in __smp_call_function_single Just got my 6 way machine to a state where cpu 0 is in an endless loop within __smp_call_function_single. All other cpus are idle. The call trace on cpu 0 looks like this: __smp_call_function_single scheduler_tick update_process_times tick_sched_timer __run_hrtimer hrtimer_interrupt clock_comparator_work do_extint ext_int_handler ----> timer irq cpu_idle __smp_call_function_single() got called from nohz_balancer_kick() (inlined) with the remote cpu being 1, wait being 0 and the per cpu variable remote_sched_softirq_cb (call_single_data) of the current cpu (0). Then it loops forever when it tries to grab the lock of the call_single_data, since it is already locked and enqueued on cpu 0. My theory how this could have happened: for some reason the scheduler decided to call __smp_call_function_single() on it's own cpu, and sends an IPI to itself. The interrupt stays pending since IRQs are disabled. If then the hypervisor schedules the cpu away it might happen that upon rescheduling both the IPI and the timer IRQ are pending. If then interrupts are enabled again it depends which one gets scheduled first. If the timer interrupt gets delivered first we end up with the local deadlock as seen in the calltrace above. Let's make __smp_call_function_single() check if the target cpu is the current cpu and execute the function immediately just like smp_call_function_single does. That should prevent at least the scenario described here. It might also be that the scheduler is not supposed to call __smp_call_function_single with the remote cpu being the current cpu, but that is a different issue. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jens Axboe <jaxboe@fusionio.com> Cc: Venkatesh Pallipadi <venki@google.com> Cc: Suresh Siddha <suresh.b.siddha@intel.com> LKML-Reference: <20100910114729.GB2827@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-10 11:47:29 +00:00
* @wait: If true, wait until function has completed on specified CPU.
*
* Like smp_call_function_single(), but allow caller to pass in a
* pre-allocated data structure. Useful for embedding @data inside
* other structures, for instance.
*/
void __smp_call_function_single(int cpu, struct call_single_data *csd,
int wait)
{
generic-ipi: Fix deadlock in __smp_call_function_single Just got my 6 way machine to a state where cpu 0 is in an endless loop within __smp_call_function_single. All other cpus are idle. The call trace on cpu 0 looks like this: __smp_call_function_single scheduler_tick update_process_times tick_sched_timer __run_hrtimer hrtimer_interrupt clock_comparator_work do_extint ext_int_handler ----> timer irq cpu_idle __smp_call_function_single() got called from nohz_balancer_kick() (inlined) with the remote cpu being 1, wait being 0 and the per cpu variable remote_sched_softirq_cb (call_single_data) of the current cpu (0). Then it loops forever when it tries to grab the lock of the call_single_data, since it is already locked and enqueued on cpu 0. My theory how this could have happened: for some reason the scheduler decided to call __smp_call_function_single() on it's own cpu, and sends an IPI to itself. The interrupt stays pending since IRQs are disabled. If then the hypervisor schedules the cpu away it might happen that upon rescheduling both the IPI and the timer IRQ are pending. If then interrupts are enabled again it depends which one gets scheduled first. If the timer interrupt gets delivered first we end up with the local deadlock as seen in the calltrace above. Let's make __smp_call_function_single() check if the target cpu is the current cpu and execute the function immediately just like smp_call_function_single does. That should prevent at least the scenario described here. It might also be that the scheduler is not supposed to call __smp_call_function_single with the remote cpu being the current cpu, but that is a different issue. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jens Axboe <jaxboe@fusionio.com> Cc: Venkatesh Pallipadi <venki@google.com> Cc: Suresh Siddha <suresh.b.siddha@intel.com> LKML-Reference: <20100910114729.GB2827@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-10 11:47:29 +00:00
unsigned int this_cpu;
unsigned long flags;
generic-ipi: Fix deadlock in __smp_call_function_single Just got my 6 way machine to a state where cpu 0 is in an endless loop within __smp_call_function_single. All other cpus are idle. The call trace on cpu 0 looks like this: __smp_call_function_single scheduler_tick update_process_times tick_sched_timer __run_hrtimer hrtimer_interrupt clock_comparator_work do_extint ext_int_handler ----> timer irq cpu_idle __smp_call_function_single() got called from nohz_balancer_kick() (inlined) with the remote cpu being 1, wait being 0 and the per cpu variable remote_sched_softirq_cb (call_single_data) of the current cpu (0). Then it loops forever when it tries to grab the lock of the call_single_data, since it is already locked and enqueued on cpu 0. My theory how this could have happened: for some reason the scheduler decided to call __smp_call_function_single() on it's own cpu, and sends an IPI to itself. The interrupt stays pending since IRQs are disabled. If then the hypervisor schedules the cpu away it might happen that upon rescheduling both the IPI and the timer IRQ are pending. If then interrupts are enabled again it depends which one gets scheduled first. If the timer interrupt gets delivered first we end up with the local deadlock as seen in the calltrace above. Let's make __smp_call_function_single() check if the target cpu is the current cpu and execute the function immediately just like smp_call_function_single does. That should prevent at least the scenario described here. It might also be that the scheduler is not supposed to call __smp_call_function_single with the remote cpu being the current cpu, but that is a different issue. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jens Axboe <jaxboe@fusionio.com> Cc: Venkatesh Pallipadi <venki@google.com> Cc: Suresh Siddha <suresh.b.siddha@intel.com> LKML-Reference: <20100910114729.GB2827@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-10 11:47:29 +00:00
this_cpu = get_cpu();
/*
* Can deadlock when called with interrupts disabled.
* We allow cpu's that are not yet online though, as no one else can
* send smp call function interrupt to this cpu and as such deadlocks
* can't happen.
*/
WARN_ON_ONCE(cpu_online(smp_processor_id()) && wait && irqs_disabled()
&& !oops_in_progress);
generic-ipi: Fix deadlock in __smp_call_function_single Just got my 6 way machine to a state where cpu 0 is in an endless loop within __smp_call_function_single. All other cpus are idle. The call trace on cpu 0 looks like this: __smp_call_function_single scheduler_tick update_process_times tick_sched_timer __run_hrtimer hrtimer_interrupt clock_comparator_work do_extint ext_int_handler ----> timer irq cpu_idle __smp_call_function_single() got called from nohz_balancer_kick() (inlined) with the remote cpu being 1, wait being 0 and the per cpu variable remote_sched_softirq_cb (call_single_data) of the current cpu (0). Then it loops forever when it tries to grab the lock of the call_single_data, since it is already locked and enqueued on cpu 0. My theory how this could have happened: for some reason the scheduler decided to call __smp_call_function_single() on it's own cpu, and sends an IPI to itself. The interrupt stays pending since IRQs are disabled. If then the hypervisor schedules the cpu away it might happen that upon rescheduling both the IPI and the timer IRQ are pending. If then interrupts are enabled again it depends which one gets scheduled first. If the timer interrupt gets delivered first we end up with the local deadlock as seen in the calltrace above. Let's make __smp_call_function_single() check if the target cpu is the current cpu and execute the function immediately just like smp_call_function_single does. That should prevent at least the scenario described here. It might also be that the scheduler is not supposed to call __smp_call_function_single with the remote cpu being the current cpu, but that is a different issue. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jens Axboe <jaxboe@fusionio.com> Cc: Venkatesh Pallipadi <venki@google.com> Cc: Suresh Siddha <suresh.b.siddha@intel.com> LKML-Reference: <20100910114729.GB2827@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-10 11:47:29 +00:00
if (cpu == this_cpu) {
local_irq_save(flags);
csd->func(csd->info);
generic-ipi: Fix deadlock in __smp_call_function_single Just got my 6 way machine to a state where cpu 0 is in an endless loop within __smp_call_function_single. All other cpus are idle. The call trace on cpu 0 looks like this: __smp_call_function_single scheduler_tick update_process_times tick_sched_timer __run_hrtimer hrtimer_interrupt clock_comparator_work do_extint ext_int_handler ----> timer irq cpu_idle __smp_call_function_single() got called from nohz_balancer_kick() (inlined) with the remote cpu being 1, wait being 0 and the per cpu variable remote_sched_softirq_cb (call_single_data) of the current cpu (0). Then it loops forever when it tries to grab the lock of the call_single_data, since it is already locked and enqueued on cpu 0. My theory how this could have happened: for some reason the scheduler decided to call __smp_call_function_single() on it's own cpu, and sends an IPI to itself. The interrupt stays pending since IRQs are disabled. If then the hypervisor schedules the cpu away it might happen that upon rescheduling both the IPI and the timer IRQ are pending. If then interrupts are enabled again it depends which one gets scheduled first. If the timer interrupt gets delivered first we end up with the local deadlock as seen in the calltrace above. Let's make __smp_call_function_single() check if the target cpu is the current cpu and execute the function immediately just like smp_call_function_single does. That should prevent at least the scenario described here. It might also be that the scheduler is not supposed to call __smp_call_function_single with the remote cpu being the current cpu, but that is a different issue. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jens Axboe <jaxboe@fusionio.com> Cc: Venkatesh Pallipadi <venki@google.com> Cc: Suresh Siddha <suresh.b.siddha@intel.com> LKML-Reference: <20100910114729.GB2827@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-10 11:47:29 +00:00
local_irq_restore(flags);
} else {
csd_lock(csd);
generic_exec_single(cpu, csd, wait);
generic-ipi: Fix deadlock in __smp_call_function_single Just got my 6 way machine to a state where cpu 0 is in an endless loop within __smp_call_function_single. All other cpus are idle. The call trace on cpu 0 looks like this: __smp_call_function_single scheduler_tick update_process_times tick_sched_timer __run_hrtimer hrtimer_interrupt clock_comparator_work do_extint ext_int_handler ----> timer irq cpu_idle __smp_call_function_single() got called from nohz_balancer_kick() (inlined) with the remote cpu being 1, wait being 0 and the per cpu variable remote_sched_softirq_cb (call_single_data) of the current cpu (0). Then it loops forever when it tries to grab the lock of the call_single_data, since it is already locked and enqueued on cpu 0. My theory how this could have happened: for some reason the scheduler decided to call __smp_call_function_single() on it's own cpu, and sends an IPI to itself. The interrupt stays pending since IRQs are disabled. If then the hypervisor schedules the cpu away it might happen that upon rescheduling both the IPI and the timer IRQ are pending. If then interrupts are enabled again it depends which one gets scheduled first. If the timer interrupt gets delivered first we end up with the local deadlock as seen in the calltrace above. Let's make __smp_call_function_single() check if the target cpu is the current cpu and execute the function immediately just like smp_call_function_single does. That should prevent at least the scenario described here. It might also be that the scheduler is not supposed to call __smp_call_function_single with the remote cpu being the current cpu, but that is a different issue. Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Jens Axboe <jaxboe@fusionio.com> Cc: Venkatesh Pallipadi <venki@google.com> Cc: Suresh Siddha <suresh.b.siddha@intel.com> LKML-Reference: <20100910114729.GB2827@osiris.boeblingen.de.ibm.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2010-09-10 11:47:29 +00:00
}
put_cpu();
}
EXPORT_SYMBOL_GPL(__smp_call_function_single);
/**
* smp_call_function_many(): Run a function on a set of other CPUs.
* @mask: The set of cpus to run on (only runs on online subset).
* @func: The function to run. This must be fast and non-blocking.
* @info: An arbitrary pointer to pass to the function.
* @wait: If true, wait (atomically) until function has completed
* on other CPUs.
*
* If @wait is true, then returns once @func has returned.
*
* You must not call this function with disabled interrupts or from a
* hardware interrupt handler or from a bottom half handler. Preemption
* must be disabled when calling this function.
*/
void smp_call_function_many(const struct cpumask *mask,
smp_call_func_t func, void *info, bool wait)
{
struct call_function_data *cfd;
smp: make smp_call_function_many() use logic similar to smp_call_function_single() I'm testing swapout workload in a two-socket Xeon machine. The workload has 10 threads, each thread sequentially accesses separate memory region. TLB flush overhead is very big in the workload. For each page, page reclaim need move it from active lru list and then unmap it. Both need a TLB flush. And this is a multthread workload, TLB flush happens in 10 CPUs. In X86, TLB flush uses generic smp_call)function. So this workload stress smp_call_function_many heavily. Without patch, perf shows: + 24.49% [k] generic_smp_call_function_interrupt - 21.72% [k] _raw_spin_lock - _raw_spin_lock + 79.80% __page_check_address + 6.42% generic_smp_call_function_interrupt + 3.31% get_swap_page + 2.37% free_pcppages_bulk + 1.75% handle_pte_fault + 1.54% put_super + 1.41% grab_super_passive + 1.36% __swap_duplicate + 0.68% blk_flush_plug_list + 0.62% swap_info_get + 6.55% [k] flush_tlb_func + 6.46% [k] smp_call_function_many + 5.09% [k] call_function_interrupt + 4.75% [k] default_send_IPI_mask_sequence_phys + 2.18% [k] find_next_bit swapout throughput is around 1300M/s. With the patch, perf shows: - 27.23% [k] _raw_spin_lock - _raw_spin_lock + 80.53% __page_check_address + 8.39% generic_smp_call_function_single_interrupt + 2.44% get_swap_page + 1.76% free_pcppages_bulk + 1.40% handle_pte_fault + 1.15% __swap_duplicate + 1.05% put_super + 0.98% grab_super_passive + 0.86% blk_flush_plug_list + 0.57% swap_info_get + 8.25% [k] default_send_IPI_mask_sequence_phys + 7.55% [k] call_function_interrupt + 7.47% [k] smp_call_function_many + 7.25% [k] flush_tlb_func + 3.81% [k] _raw_spin_lock_irqsave + 3.78% [k] generic_smp_call_function_single_interrupt swapout throughput is around 1400M/s. So there is around a 7% improvement, and total cpu utilization doesn't change. Without the patch, cfd_data is shared by all CPUs. generic_smp_call_function_interrupt does read/write cfd_data several times which will create a lot of cache ping-pong. With the patch, the data becomes per-cpu. The ping-pong is avoided. And from the perf data, this doesn't make call_single_queue lock contend. Next step is to remove generic_smp_call_function_interrupt() from arch code. Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 00:43:03 +00:00
int cpu, next_cpu, this_cpu = smp_processor_id();
/*
* Can deadlock when called with interrupts disabled.
* We allow cpu's that are not yet online though, as no one else can
* send smp call function interrupt to this cpu and as such deadlocks
* can't happen.
*/
WARN_ON_ONCE(cpu_online(this_cpu) && irqs_disabled()
&& !oops_in_progress && !early_boot_irqs_disabled);
/* Try to fastpath. So, what's a CPU they want? Ignoring this one. */
cpu = cpumask_first_and(mask, cpu_online_mask);
if (cpu == this_cpu)
cpu = cpumask_next_and(cpu, mask, cpu_online_mask);
/* No online cpus? We're done. */
if (cpu >= nr_cpu_ids)
return;
/* Do we have another CPU which isn't us? */
next_cpu = cpumask_next_and(cpu, mask, cpu_online_mask);
if (next_cpu == this_cpu)
next_cpu = cpumask_next_and(next_cpu, mask, cpu_online_mask);
/* Fastpath: do that cpu by itself. */
if (next_cpu >= nr_cpu_ids) {
smp_call_function_single(cpu, func, info, wait);
return;
}
cfd = &__get_cpu_var(cfd_data);
call_function_many: add missing ordering Paul McKenney's review pointed out two problems with the barriers in the 2.6.38 update to the smp call function many code. First, a barrier that would force the func and info members of data to be visible before their consumption in the interrupt handler was missing. This can be solved by adding a smp_wmb between setting the func and info members and setting setting the cpumask; this will pair with the existing and required smp_rmb ordering the cpumask read before the read of refs. This placement avoids the need a second smp_rmb in the interrupt handler which would be executed on each of the N cpus executing the call request. (I was thinking this barrier was present but was not). Second, the previous write to refs (establishing the zero that we the interrupt handler was testing from all cpus) was performed by a third party cpu. This would invoke transitivity which, as a recient or concurrent addition to memory-barriers.txt now explicitly states, would require a full smp_mb(). However, we know the cpumask will only be set by one cpu (the data owner) and any preivous iteration of the mask would have cleared by the reading cpu. By redundantly writing refs to 0 on the owning cpu before the smp_wmb, the write to refs will follow the same path as the writes that set the cpumask, which in turn allows us to keep the barrier in the interrupt handler a smp_rmb instead of promoting it to a smp_mb (which will be be executed by N cpus for each of the possible M elements on the list). I moved and expanded the comment about our (ab)use of the rcu list primitives for the concurrent walk earlier into this function. I considered moving the first two paragraphs to the queue list head and lock, but felt it would have been too disconected from the code. Cc: Paul McKinney <paulmck@linux.vnet.ibm.com> Cc: stable@kernel.org (2.6.32 and later) Signed-off-by: Milton Miller <miltonm@bga.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-03-15 19:27:16 +00:00
cpumask_and(cfd->cpumask, mask, cpu_online_mask);
cpumask_clear_cpu(this_cpu, cfd->cpumask);
/* Some callers race with other cpus changing the passed mask */
if (unlikely(!cpumask_weight(cfd->cpumask)))
return;
for_each_cpu(cpu, cfd->cpumask) {
struct call_single_data *csd = per_cpu_ptr(cfd->csd, cpu);
smp: make smp_call_function_many() use logic similar to smp_call_function_single() I'm testing swapout workload in a two-socket Xeon machine. The workload has 10 threads, each thread sequentially accesses separate memory region. TLB flush overhead is very big in the workload. For each page, page reclaim need move it from active lru list and then unmap it. Both need a TLB flush. And this is a multthread workload, TLB flush happens in 10 CPUs. In X86, TLB flush uses generic smp_call)function. So this workload stress smp_call_function_many heavily. Without patch, perf shows: + 24.49% [k] generic_smp_call_function_interrupt - 21.72% [k] _raw_spin_lock - _raw_spin_lock + 79.80% __page_check_address + 6.42% generic_smp_call_function_interrupt + 3.31% get_swap_page + 2.37% free_pcppages_bulk + 1.75% handle_pte_fault + 1.54% put_super + 1.41% grab_super_passive + 1.36% __swap_duplicate + 0.68% blk_flush_plug_list + 0.62% swap_info_get + 6.55% [k] flush_tlb_func + 6.46% [k] smp_call_function_many + 5.09% [k] call_function_interrupt + 4.75% [k] default_send_IPI_mask_sequence_phys + 2.18% [k] find_next_bit swapout throughput is around 1300M/s. With the patch, perf shows: - 27.23% [k] _raw_spin_lock - _raw_spin_lock + 80.53% __page_check_address + 8.39% generic_smp_call_function_single_interrupt + 2.44% get_swap_page + 1.76% free_pcppages_bulk + 1.40% handle_pte_fault + 1.15% __swap_duplicate + 1.05% put_super + 0.98% grab_super_passive + 0.86% blk_flush_plug_list + 0.57% swap_info_get + 8.25% [k] default_send_IPI_mask_sequence_phys + 7.55% [k] call_function_interrupt + 7.47% [k] smp_call_function_many + 7.25% [k] flush_tlb_func + 3.81% [k] _raw_spin_lock_irqsave + 3.78% [k] generic_smp_call_function_single_interrupt swapout throughput is around 1400M/s. So there is around a 7% improvement, and total cpu utilization doesn't change. Without the patch, cfd_data is shared by all CPUs. generic_smp_call_function_interrupt does read/write cfd_data several times which will create a lot of cache ping-pong. With the patch, the data becomes per-cpu. The ping-pong is avoided. And from the perf data, this doesn't make call_single_queue lock contend. Next step is to remove generic_smp_call_function_interrupt() from arch code. Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 00:43:03 +00:00
csd_lock(csd);
csd->func = func;
csd->info = info;
llist_add(&csd->llist, &per_cpu(call_single_queue, cpu));
smp: make smp_call_function_many() use logic similar to smp_call_function_single() I'm testing swapout workload in a two-socket Xeon machine. The workload has 10 threads, each thread sequentially accesses separate memory region. TLB flush overhead is very big in the workload. For each page, page reclaim need move it from active lru list and then unmap it. Both need a TLB flush. And this is a multthread workload, TLB flush happens in 10 CPUs. In X86, TLB flush uses generic smp_call)function. So this workload stress smp_call_function_many heavily. Without patch, perf shows: + 24.49% [k] generic_smp_call_function_interrupt - 21.72% [k] _raw_spin_lock - _raw_spin_lock + 79.80% __page_check_address + 6.42% generic_smp_call_function_interrupt + 3.31% get_swap_page + 2.37% free_pcppages_bulk + 1.75% handle_pte_fault + 1.54% put_super + 1.41% grab_super_passive + 1.36% __swap_duplicate + 0.68% blk_flush_plug_list + 0.62% swap_info_get + 6.55% [k] flush_tlb_func + 6.46% [k] smp_call_function_many + 5.09% [k] call_function_interrupt + 4.75% [k] default_send_IPI_mask_sequence_phys + 2.18% [k] find_next_bit swapout throughput is around 1300M/s. With the patch, perf shows: - 27.23% [k] _raw_spin_lock - _raw_spin_lock + 80.53% __page_check_address + 8.39% generic_smp_call_function_single_interrupt + 2.44% get_swap_page + 1.76% free_pcppages_bulk + 1.40% handle_pte_fault + 1.15% __swap_duplicate + 1.05% put_super + 0.98% grab_super_passive + 0.86% blk_flush_plug_list + 0.57% swap_info_get + 8.25% [k] default_send_IPI_mask_sequence_phys + 7.55% [k] call_function_interrupt + 7.47% [k] smp_call_function_many + 7.25% [k] flush_tlb_func + 3.81% [k] _raw_spin_lock_irqsave + 3.78% [k] generic_smp_call_function_single_interrupt swapout throughput is around 1400M/s. So there is around a 7% improvement, and total cpu utilization doesn't change. Without the patch, cfd_data is shared by all CPUs. generic_smp_call_function_interrupt does read/write cfd_data several times which will create a lot of cache ping-pong. With the patch, the data becomes per-cpu. The ping-pong is avoided. And from the perf data, this doesn't make call_single_queue lock contend. Next step is to remove generic_smp_call_function_interrupt() from arch code. Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 00:43:03 +00:00
}
/* Send a message to all CPUs in the map */
arch_send_call_function_ipi_mask(cfd->cpumask);
smp: make smp_call_function_many() use logic similar to smp_call_function_single() I'm testing swapout workload in a two-socket Xeon machine. The workload has 10 threads, each thread sequentially accesses separate memory region. TLB flush overhead is very big in the workload. For each page, page reclaim need move it from active lru list and then unmap it. Both need a TLB flush. And this is a multthread workload, TLB flush happens in 10 CPUs. In X86, TLB flush uses generic smp_call)function. So this workload stress smp_call_function_many heavily. Without patch, perf shows: + 24.49% [k] generic_smp_call_function_interrupt - 21.72% [k] _raw_spin_lock - _raw_spin_lock + 79.80% __page_check_address + 6.42% generic_smp_call_function_interrupt + 3.31% get_swap_page + 2.37% free_pcppages_bulk + 1.75% handle_pte_fault + 1.54% put_super + 1.41% grab_super_passive + 1.36% __swap_duplicate + 0.68% blk_flush_plug_list + 0.62% swap_info_get + 6.55% [k] flush_tlb_func + 6.46% [k] smp_call_function_many + 5.09% [k] call_function_interrupt + 4.75% [k] default_send_IPI_mask_sequence_phys + 2.18% [k] find_next_bit swapout throughput is around 1300M/s. With the patch, perf shows: - 27.23% [k] _raw_spin_lock - _raw_spin_lock + 80.53% __page_check_address + 8.39% generic_smp_call_function_single_interrupt + 2.44% get_swap_page + 1.76% free_pcppages_bulk + 1.40% handle_pte_fault + 1.15% __swap_duplicate + 1.05% put_super + 0.98% grab_super_passive + 0.86% blk_flush_plug_list + 0.57% swap_info_get + 8.25% [k] default_send_IPI_mask_sequence_phys + 7.55% [k] call_function_interrupt + 7.47% [k] smp_call_function_many + 7.25% [k] flush_tlb_func + 3.81% [k] _raw_spin_lock_irqsave + 3.78% [k] generic_smp_call_function_single_interrupt swapout throughput is around 1400M/s. So there is around a 7% improvement, and total cpu utilization doesn't change. Without the patch, cfd_data is shared by all CPUs. generic_smp_call_function_interrupt does read/write cfd_data several times which will create a lot of cache ping-pong. With the patch, the data becomes per-cpu. The ping-pong is avoided. And from the perf data, this doesn't make call_single_queue lock contend. Next step is to remove generic_smp_call_function_interrupt() from arch code. Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 00:43:03 +00:00
if (wait) {
for_each_cpu(cpu, cfd->cpumask) {
struct call_single_data *csd;
csd = per_cpu_ptr(cfd->csd, cpu);
smp: make smp_call_function_many() use logic similar to smp_call_function_single() I'm testing swapout workload in a two-socket Xeon machine. The workload has 10 threads, each thread sequentially accesses separate memory region. TLB flush overhead is very big in the workload. For each page, page reclaim need move it from active lru list and then unmap it. Both need a TLB flush. And this is a multthread workload, TLB flush happens in 10 CPUs. In X86, TLB flush uses generic smp_call)function. So this workload stress smp_call_function_many heavily. Without patch, perf shows: + 24.49% [k] generic_smp_call_function_interrupt - 21.72% [k] _raw_spin_lock - _raw_spin_lock + 79.80% __page_check_address + 6.42% generic_smp_call_function_interrupt + 3.31% get_swap_page + 2.37% free_pcppages_bulk + 1.75% handle_pte_fault + 1.54% put_super + 1.41% grab_super_passive + 1.36% __swap_duplicate + 0.68% blk_flush_plug_list + 0.62% swap_info_get + 6.55% [k] flush_tlb_func + 6.46% [k] smp_call_function_many + 5.09% [k] call_function_interrupt + 4.75% [k] default_send_IPI_mask_sequence_phys + 2.18% [k] find_next_bit swapout throughput is around 1300M/s. With the patch, perf shows: - 27.23% [k] _raw_spin_lock - _raw_spin_lock + 80.53% __page_check_address + 8.39% generic_smp_call_function_single_interrupt + 2.44% get_swap_page + 1.76% free_pcppages_bulk + 1.40% handle_pte_fault + 1.15% __swap_duplicate + 1.05% put_super + 0.98% grab_super_passive + 0.86% blk_flush_plug_list + 0.57% swap_info_get + 8.25% [k] default_send_IPI_mask_sequence_phys + 7.55% [k] call_function_interrupt + 7.47% [k] smp_call_function_many + 7.25% [k] flush_tlb_func + 3.81% [k] _raw_spin_lock_irqsave + 3.78% [k] generic_smp_call_function_single_interrupt swapout throughput is around 1400M/s. So there is around a 7% improvement, and total cpu utilization doesn't change. Without the patch, cfd_data is shared by all CPUs. generic_smp_call_function_interrupt does read/write cfd_data several times which will create a lot of cache ping-pong. With the patch, the data becomes per-cpu. The ping-pong is avoided. And from the perf data, this doesn't make call_single_queue lock contend. Next step is to remove generic_smp_call_function_interrupt() from arch code. Signed-off-by: Shaohua Li <shli@fusionio.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Ingo Molnar <mingo@elte.hu> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-22 00:43:03 +00:00
csd_lock_wait(csd);
}
}
}
EXPORT_SYMBOL(smp_call_function_many);
/**
* smp_call_function(): Run a function on all other CPUs.
* @func: The function to run. This must be fast and non-blocking.
* @info: An arbitrary pointer to pass to the function.
* @wait: If true, wait (atomically) until function has completed
* on other CPUs.
*
* Returns 0.
*
* If @wait is true, then returns once @func has returned; otherwise
* it returns just before the target cpu calls @func.
*
* You must not call this function with disabled interrupts or from a
* hardware interrupt handler or from a bottom half handler.
*/
int smp_call_function(smp_call_func_t func, void *info, int wait)
{
preempt_disable();
smp_call_function_many(cpu_online_mask, func, info, wait);
preempt_enable();
return 0;
}
EXPORT_SYMBOL(smp_call_function);
/* Setup configured maximum number of CPUs to activate */
unsigned int setup_max_cpus = NR_CPUS;
EXPORT_SYMBOL(setup_max_cpus);
/*
* Setup routine for controlling SMP activation
*
* Command-line option of "nosmp" or "maxcpus=0" will disable SMP
* activation entirely (the MPS table probe still happens, though).
*
* Command-line option of "maxcpus=<NUM>", where <NUM> is an integer
* greater than 0, limits the maximum number of CPUs activated in
* SMP mode to <NUM>.
*/
void __weak arch_disable_smp_support(void) { }
static int __init nosmp(char *str)
{
setup_max_cpus = 0;
arch_disable_smp_support();
return 0;
}
early_param("nosmp", nosmp);
/* this is hard limit */
static int __init nrcpus(char *str)
{
int nr_cpus;
get_option(&str, &nr_cpus);
if (nr_cpus > 0 && nr_cpus < nr_cpu_ids)
nr_cpu_ids = nr_cpus;
return 0;
}
early_param("nr_cpus", nrcpus);
static int __init maxcpus(char *str)
{
get_option(&str, &setup_max_cpus);
if (setup_max_cpus == 0)
arch_disable_smp_support();
return 0;
}
early_param("maxcpus", maxcpus);
/* Setup number of possible processor ids */
int nr_cpu_ids __read_mostly = NR_CPUS;
EXPORT_SYMBOL(nr_cpu_ids);
/* An arch may set nr_cpu_ids earlier if needed, so this would be redundant */
void __init setup_nr_cpu_ids(void)
{
nr_cpu_ids = find_last_bit(cpumask_bits(cpu_possible_mask),NR_CPUS) + 1;
}
x86/boot: Further compress CPUs bootup message Turn it into (for example): [ 0.073380] x86: Booting SMP configuration: [ 0.074005] .... node #0, CPUs: #1 #2 #3 #4 #5 #6 #7 [ 0.603005] .... node #1, CPUs: #8 #9 #10 #11 #12 #13 #14 #15 [ 1.200005] .... node #2, CPUs: #16 #17 #18 #19 #20 #21 #22 #23 [ 1.796005] .... node #3, CPUs: #24 #25 #26 #27 #28 #29 #30 #31 [ 2.393005] .... node #4, CPUs: #32 #33 #34 #35 #36 #37 #38 #39 [ 2.996005] .... node #5, CPUs: #40 #41 #42 #43 #44 #45 #46 #47 [ 3.600005] .... node #6, CPUs: #48 #49 #50 #51 #52 #53 #54 #55 [ 4.202005] .... node #7, CPUs: #56 #57 #58 #59 #60 #61 #62 #63 [ 4.811005] .... node #8, CPUs: #64 #65 #66 #67 #68 #69 #70 #71 [ 5.421006] .... node #9, CPUs: #72 #73 #74 #75 #76 #77 #78 #79 [ 6.032005] .... node #10, CPUs: #80 #81 #82 #83 #84 #85 #86 #87 [ 6.648006] .... node #11, CPUs: #88 #89 #90 #91 #92 #93 #94 #95 [ 7.262005] .... node #12, CPUs: #96 #97 #98 #99 #100 #101 #102 #103 [ 7.865005] .... node #13, CPUs: #104 #105 #106 #107 #108 #109 #110 #111 [ 8.466005] .... node #14, CPUs: #112 #113 #114 #115 #116 #117 #118 #119 [ 9.073006] .... node #15, CPUs: #120 #121 #122 #123 #124 #125 #126 #127 [ 9.679901] x86: Booted up 16 nodes, 128 CPUs and drop useless elements. Change num_digits() to hpa's division-avoiding, cell-phone-typed version which he went at great lengths and pains to submit on a Saturday evening. Signed-off-by: Borislav Petkov <bp@suse.de> Cc: huawei.libin@huawei.com Cc: wangyijing@huawei.com Cc: fenghua.yu@intel.com Cc: guohanjun@huawei.com Cc: paul.gortmaker@windriver.com Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/20130930095624.GB16383@pd.tnic Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-09-30 09:56:24 +00:00
void __weak smp_announce(void)
{
printk(KERN_INFO "Brought up %d CPUs\n", num_online_cpus());
}
/* Called by boot processor to activate the rest. */
void __init smp_init(void)
{
unsigned int cpu;
idle_threads_init();
/* FIXME: This should be done in userspace --RR */
for_each_present_cpu(cpu) {
if (num_online_cpus() >= setup_max_cpus)
break;
if (!cpu_online(cpu))
cpu_up(cpu);
}
/* Any cleanup work */
x86/boot: Further compress CPUs bootup message Turn it into (for example): [ 0.073380] x86: Booting SMP configuration: [ 0.074005] .... node #0, CPUs: #1 #2 #3 #4 #5 #6 #7 [ 0.603005] .... node #1, CPUs: #8 #9 #10 #11 #12 #13 #14 #15 [ 1.200005] .... node #2, CPUs: #16 #17 #18 #19 #20 #21 #22 #23 [ 1.796005] .... node #3, CPUs: #24 #25 #26 #27 #28 #29 #30 #31 [ 2.393005] .... node #4, CPUs: #32 #33 #34 #35 #36 #37 #38 #39 [ 2.996005] .... node #5, CPUs: #40 #41 #42 #43 #44 #45 #46 #47 [ 3.600005] .... node #6, CPUs: #48 #49 #50 #51 #52 #53 #54 #55 [ 4.202005] .... node #7, CPUs: #56 #57 #58 #59 #60 #61 #62 #63 [ 4.811005] .... node #8, CPUs: #64 #65 #66 #67 #68 #69 #70 #71 [ 5.421006] .... node #9, CPUs: #72 #73 #74 #75 #76 #77 #78 #79 [ 6.032005] .... node #10, CPUs: #80 #81 #82 #83 #84 #85 #86 #87 [ 6.648006] .... node #11, CPUs: #88 #89 #90 #91 #92 #93 #94 #95 [ 7.262005] .... node #12, CPUs: #96 #97 #98 #99 #100 #101 #102 #103 [ 7.865005] .... node #13, CPUs: #104 #105 #106 #107 #108 #109 #110 #111 [ 8.466005] .... node #14, CPUs: #112 #113 #114 #115 #116 #117 #118 #119 [ 9.073006] .... node #15, CPUs: #120 #121 #122 #123 #124 #125 #126 #127 [ 9.679901] x86: Booted up 16 nodes, 128 CPUs and drop useless elements. Change num_digits() to hpa's division-avoiding, cell-phone-typed version which he went at great lengths and pains to submit on a Saturday evening. Signed-off-by: Borislav Petkov <bp@suse.de> Cc: huawei.libin@huawei.com Cc: wangyijing@huawei.com Cc: fenghua.yu@intel.com Cc: guohanjun@huawei.com Cc: paul.gortmaker@windriver.com Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/20130930095624.GB16383@pd.tnic Signed-off-by: Ingo Molnar <mingo@kernel.org>
2013-09-30 09:56:24 +00:00
smp_announce();
smp_cpus_done(setup_max_cpus);
}
/*
* Call a function on all processors. May be used during early boot while
* early_boot_irqs_disabled is set. Use local_irq_save/restore() instead
* of local_irq_disable/enable().
*/
int on_each_cpu(void (*func) (void *info), void *info, int wait)
{
unsigned long flags;
int ret = 0;
preempt_disable();
ret = smp_call_function(func, info, wait);
local_irq_save(flags);
func(info);
local_irq_restore(flags);
preempt_enable();
return ret;
}
EXPORT_SYMBOL(on_each_cpu);
smp: introduce a generic on_each_cpu_mask() function We have lots of infrastructure in place to partition multi-core systems such that we have a group of CPUs that are dedicated to specific task: cgroups, scheduler and interrupt affinity, and cpuisol= boot parameter. Still, kernel code will at times interrupt all CPUs in the system via IPIs for various needs. These IPIs are useful and cannot be avoided altogether, but in certain cases it is possible to interrupt only specific CPUs that have useful work to do and not the entire system. This patch set, inspired by discussions with Peter Zijlstra and Frederic Weisbecker when testing the nohz task patch set, is a first stab at trying to explore doing this by locating the places where such global IPI calls are being made and turning the global IPI into an IPI for a specific group of CPUs. The purpose of the patch set is to get feedback if this is the right way to go for dealing with this issue and indeed, if the issue is even worth dealing with at all. Based on the feedback from this patch set I plan to offer further patches that address similar issue in other code paths. This patch creates an on_each_cpu_mask() and on_each_cpu_cond() infrastructure API (the former derived from existing arch specific versions in Tile and Arm) and uses them to turn several global IPI invocation to per CPU group invocations. Core kernel: on_each_cpu_mask() calls a function on processors specified by cpumask, which may or may not include the local processor. You must not call this function with disabled interrupts or from a hardware interrupt handler or from a bottom half handler. arch/arm: Note that the generic version is a little different then the Arm one: 1. It has the mask as first parameter 2. It calls the function on the calling CPU with interrupts disabled, but this should be OK since the function is called on the other CPUs with interrupts disabled anyway. arch/tile: The API is the same as the tile private one, but the generic version also calls the function on the with interrupts disabled in UP case This is OK since the function is called on the other CPUs with interrupts disabled. Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com> Reviewed-by: Christoph Lameter <cl@linux.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Pekka Enberg <penberg@kernel.org> Cc: Matt Mackall <mpm@selenic.com> Cc: Rik van Riel <riel@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Sasha Levin <levinsasha928@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Avi Kivity <avi@redhat.com> Acked-by: Michal Nazarewicz <mina86@mina86.org> Cc: Kosaki Motohiro <kosaki.motohiro@gmail.com> Cc: Milton Miller <miltonm@bga.com> Cc: Russell King <linux@arm.linux.org.uk> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-28 21:42:43 +00:00
/**
* on_each_cpu_mask(): Run a function on processors specified by
* cpumask, which may include the local processor.
* @mask: The set of cpus to run on (only runs on online subset).
* @func: The function to run. This must be fast and non-blocking.
* @info: An arbitrary pointer to pass to the function.
* @wait: If true, wait (atomically) until function has completed
* on other CPUs.
*
* If @wait is true, then returns once @func has returned.
*
* You must not call this function with disabled interrupts or from a
* hardware interrupt handler or from a bottom half handler. The
* exception is that it may be used during early boot while
* early_boot_irqs_disabled is set.
smp: introduce a generic on_each_cpu_mask() function We have lots of infrastructure in place to partition multi-core systems such that we have a group of CPUs that are dedicated to specific task: cgroups, scheduler and interrupt affinity, and cpuisol= boot parameter. Still, kernel code will at times interrupt all CPUs in the system via IPIs for various needs. These IPIs are useful and cannot be avoided altogether, but in certain cases it is possible to interrupt only specific CPUs that have useful work to do and not the entire system. This patch set, inspired by discussions with Peter Zijlstra and Frederic Weisbecker when testing the nohz task patch set, is a first stab at trying to explore doing this by locating the places where such global IPI calls are being made and turning the global IPI into an IPI for a specific group of CPUs. The purpose of the patch set is to get feedback if this is the right way to go for dealing with this issue and indeed, if the issue is even worth dealing with at all. Based on the feedback from this patch set I plan to offer further patches that address similar issue in other code paths. This patch creates an on_each_cpu_mask() and on_each_cpu_cond() infrastructure API (the former derived from existing arch specific versions in Tile and Arm) and uses them to turn several global IPI invocation to per CPU group invocations. Core kernel: on_each_cpu_mask() calls a function on processors specified by cpumask, which may or may not include the local processor. You must not call this function with disabled interrupts or from a hardware interrupt handler or from a bottom half handler. arch/arm: Note that the generic version is a little different then the Arm one: 1. It has the mask as first parameter 2. It calls the function on the calling CPU with interrupts disabled, but this should be OK since the function is called on the other CPUs with interrupts disabled anyway. arch/tile: The API is the same as the tile private one, but the generic version also calls the function on the with interrupts disabled in UP case This is OK since the function is called on the other CPUs with interrupts disabled. Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com> Reviewed-by: Christoph Lameter <cl@linux.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Pekka Enberg <penberg@kernel.org> Cc: Matt Mackall <mpm@selenic.com> Cc: Rik van Riel <riel@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Sasha Levin <levinsasha928@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Avi Kivity <avi@redhat.com> Acked-by: Michal Nazarewicz <mina86@mina86.org> Cc: Kosaki Motohiro <kosaki.motohiro@gmail.com> Cc: Milton Miller <miltonm@bga.com> Cc: Russell King <linux@arm.linux.org.uk> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-28 21:42:43 +00:00
*/
void on_each_cpu_mask(const struct cpumask *mask, smp_call_func_t func,
void *info, bool wait)
{
int cpu = get_cpu();
smp_call_function_many(mask, func, info, wait);
if (cpumask_test_cpu(cpu, mask)) {
unsigned long flags;
local_irq_save(flags);
smp: introduce a generic on_each_cpu_mask() function We have lots of infrastructure in place to partition multi-core systems such that we have a group of CPUs that are dedicated to specific task: cgroups, scheduler and interrupt affinity, and cpuisol= boot parameter. Still, kernel code will at times interrupt all CPUs in the system via IPIs for various needs. These IPIs are useful and cannot be avoided altogether, but in certain cases it is possible to interrupt only specific CPUs that have useful work to do and not the entire system. This patch set, inspired by discussions with Peter Zijlstra and Frederic Weisbecker when testing the nohz task patch set, is a first stab at trying to explore doing this by locating the places where such global IPI calls are being made and turning the global IPI into an IPI for a specific group of CPUs. The purpose of the patch set is to get feedback if this is the right way to go for dealing with this issue and indeed, if the issue is even worth dealing with at all. Based on the feedback from this patch set I plan to offer further patches that address similar issue in other code paths. This patch creates an on_each_cpu_mask() and on_each_cpu_cond() infrastructure API (the former derived from existing arch specific versions in Tile and Arm) and uses them to turn several global IPI invocation to per CPU group invocations. Core kernel: on_each_cpu_mask() calls a function on processors specified by cpumask, which may or may not include the local processor. You must not call this function with disabled interrupts or from a hardware interrupt handler or from a bottom half handler. arch/arm: Note that the generic version is a little different then the Arm one: 1. It has the mask as first parameter 2. It calls the function on the calling CPU with interrupts disabled, but this should be OK since the function is called on the other CPUs with interrupts disabled anyway. arch/tile: The API is the same as the tile private one, but the generic version also calls the function on the with interrupts disabled in UP case This is OK since the function is called on the other CPUs with interrupts disabled. Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com> Reviewed-by: Christoph Lameter <cl@linux.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Pekka Enberg <penberg@kernel.org> Cc: Matt Mackall <mpm@selenic.com> Cc: Rik van Riel <riel@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Sasha Levin <levinsasha928@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Avi Kivity <avi@redhat.com> Acked-by: Michal Nazarewicz <mina86@mina86.org> Cc: Kosaki Motohiro <kosaki.motohiro@gmail.com> Cc: Milton Miller <miltonm@bga.com> Cc: Russell King <linux@arm.linux.org.uk> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-28 21:42:43 +00:00
func(info);
local_irq_restore(flags);
smp: introduce a generic on_each_cpu_mask() function We have lots of infrastructure in place to partition multi-core systems such that we have a group of CPUs that are dedicated to specific task: cgroups, scheduler and interrupt affinity, and cpuisol= boot parameter. Still, kernel code will at times interrupt all CPUs in the system via IPIs for various needs. These IPIs are useful and cannot be avoided altogether, but in certain cases it is possible to interrupt only specific CPUs that have useful work to do and not the entire system. This patch set, inspired by discussions with Peter Zijlstra and Frederic Weisbecker when testing the nohz task patch set, is a first stab at trying to explore doing this by locating the places where such global IPI calls are being made and turning the global IPI into an IPI for a specific group of CPUs. The purpose of the patch set is to get feedback if this is the right way to go for dealing with this issue and indeed, if the issue is even worth dealing with at all. Based on the feedback from this patch set I plan to offer further patches that address similar issue in other code paths. This patch creates an on_each_cpu_mask() and on_each_cpu_cond() infrastructure API (the former derived from existing arch specific versions in Tile and Arm) and uses them to turn several global IPI invocation to per CPU group invocations. Core kernel: on_each_cpu_mask() calls a function on processors specified by cpumask, which may or may not include the local processor. You must not call this function with disabled interrupts or from a hardware interrupt handler or from a bottom half handler. arch/arm: Note that the generic version is a little different then the Arm one: 1. It has the mask as first parameter 2. It calls the function on the calling CPU with interrupts disabled, but this should be OK since the function is called on the other CPUs with interrupts disabled anyway. arch/tile: The API is the same as the tile private one, but the generic version also calls the function on the with interrupts disabled in UP case This is OK since the function is called on the other CPUs with interrupts disabled. Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com> Reviewed-by: Christoph Lameter <cl@linux.com> Acked-by: Chris Metcalf <cmetcalf@tilera.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Pekka Enberg <penberg@kernel.org> Cc: Matt Mackall <mpm@selenic.com> Cc: Rik van Riel <riel@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Sasha Levin <levinsasha928@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Avi Kivity <avi@redhat.com> Acked-by: Michal Nazarewicz <mina86@mina86.org> Cc: Kosaki Motohiro <kosaki.motohiro@gmail.com> Cc: Milton Miller <miltonm@bga.com> Cc: Russell King <linux@arm.linux.org.uk> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-28 21:42:43 +00:00
}
put_cpu();
}
EXPORT_SYMBOL(on_each_cpu_mask);
smp: add func to IPI cpus based on parameter func Add the on_each_cpu_cond() function that wraps on_each_cpu_mask() and calculates the cpumask of cpus to IPI by calling a function supplied as a parameter in order to determine whether to IPI each specific cpu. The function works around allocation failure of cpumask variable in CONFIG_CPUMASK_OFFSTACK=y by itereating over cpus sending an IPI a time via smp_call_function_single(). The function is useful since it allows to seperate the specific code that decided in each case whether to IPI a specific cpu for a specific request from the common boilerplate code of handling creating the mask, handling failures etc. [akpm@linux-foundation.org: s/gfpflags/gfp_flags/] [akpm@linux-foundation.org: avoid double-evaluation of `info' (per Michal), parenthesise evaluation of `cond_func'] [akpm@linux-foundation.org: s/CPU/CPUs, use all 80 cols in comment] Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com> Cc: Chris Metcalf <cmetcalf@tilera.com> Cc: Christoph Lameter <cl@linux-foundation.org> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Russell King <linux@arm.linux.org.uk> Cc: Pekka Enberg <penberg@kernel.org> Cc: Matt Mackall <mpm@selenic.com> Cc: Sasha Levin <levinsasha928@gmail.com> Cc: Rik van Riel <riel@redhat.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Avi Kivity <avi@redhat.com> Acked-by: Michal Nazarewicz <mina86@mina86.org> Cc: Kosaki Motohiro <kosaki.motohiro@gmail.com> Cc: Milton Miller <miltonm@bga.com> Reviewed-by: "Srivatsa S. Bhat" <srivatsa.bhat@linux.vnet.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-28 21:42:43 +00:00
/*
* on_each_cpu_cond(): Call a function on each processor for which
* the supplied function cond_func returns true, optionally waiting
* for all the required CPUs to finish. This may include the local
* processor.
* @cond_func: A callback function that is passed a cpu id and
* the the info parameter. The function is called
* with preemption disabled. The function should
* return a blooean value indicating whether to IPI
* the specified CPU.
* @func: The function to run on all applicable CPUs.
* This must be fast and non-blocking.
* @info: An arbitrary pointer to pass to both functions.
* @wait: If true, wait (atomically) until function has
* completed on other CPUs.
* @gfp_flags: GFP flags to use when allocating the cpumask
* used internally by the function.
*
* The function might sleep if the GFP flags indicates a non
* atomic allocation is allowed.
*
* Preemption is disabled to protect against CPUs going offline but not online.
* CPUs going online during the call will not be seen or sent an IPI.
*
* You must not call this function with disabled interrupts or
* from a hardware interrupt handler or from a bottom half handler.
*/
void on_each_cpu_cond(bool (*cond_func)(int cpu, void *info),
smp_call_func_t func, void *info, bool wait,
gfp_t gfp_flags)
{
cpumask_var_t cpus;
int cpu, ret;
might_sleep_if(gfp_flags & __GFP_WAIT);
if (likely(zalloc_cpumask_var(&cpus, (gfp_flags|__GFP_NOWARN)))) {
preempt_disable();
for_each_online_cpu(cpu)
if (cond_func(cpu, info))
cpumask_set_cpu(cpu, cpus);
on_each_cpu_mask(cpus, func, info, wait);
preempt_enable();
free_cpumask_var(cpus);
} else {
/*
* No free cpumask, bother. No matter, we'll
* just have to IPI them one by one.
*/
preempt_disable();
for_each_online_cpu(cpu)
if (cond_func(cpu, info)) {
ret = smp_call_function_single(cpu, func,
info, wait);
WARN_ON_ONCE(!ret);
}
preempt_enable();
}
}
EXPORT_SYMBOL(on_each_cpu_cond);
static void do_nothing(void *unused)
{
}
/**
* kick_all_cpus_sync - Force all cpus out of idle
*
* Used to synchronize the update of pm_idle function pointer. It's
* called after the pointer is updated and returns after the dummy
* callback function has been executed on all cpus. The execution of
* the function can only happen on the remote cpus after they have
* left the idle function which had been called via pm_idle function
* pointer. So it's guaranteed that nothing uses the previous pointer
* anymore.
*/
void kick_all_cpus_sync(void)
{
/* Make sure the change is visible before we kick the cpus */
smp_mb();
smp_call_function(do_nothing, NULL, 1);
}
EXPORT_SYMBOL_GPL(kick_all_cpus_sync);