* [PATCH 0/2] minor NUMA cleanups & documentation
@ 2012-10-18 21:18 Rik van Riel
2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel
` (2 more replies)
0 siblings, 3 replies; 20+ messages in thread
From: Rik van Riel @ 2012-10-18 21:18 UTC (permalink / raw)
To: mingo; +Cc: linux-kernel, aarcange, a.p.zijlstra
Hi Ingo,
Here are some minor NUMA cleanups to start with.
I have some ideas for larger improvements and ideas to port over
from autonuma, but I got caught up in some of the code and am
not sure about those changes yet.
--
All Rights Reversed
^ permalink raw reply [flat|nested] 20+ messages in thread* [PATCH 1/2] add credits for NUMA placement 2012-10-18 21:18 [PATCH 0/2] minor NUMA cleanups & documentation Rik van Riel @ 2012-10-18 21:19 ` Rik van Riel 2012-10-19 11:41 ` Peter Zijlstra ` (5 more replies) 2012-10-18 21:20 ` [PATCH 2/2] rename NUMA fault handling functions Rik van Riel 2012-10-21 12:43 ` [PATCH 3/2] sched, numa, mm: Implement constant rate working set sampling Ingo Molnar 2 siblings, 6 replies; 20+ messages in thread From: Rik van Riel @ 2012-10-18 21:19 UTC (permalink / raw) To: mingo; +Cc: linux-kernel, aarcange, a.p.zijlstra The NUMA placement code has been rewritten several times, but the basic ideas took a lot of work to develop. The people who put in the work deserve credit for it. Thanks Andrea & Peter :) The Documentation/scheduler/numa-problem.txt file should probably be rewritten once we figure out the final details of what the NUMA code needs to do, and why. Signed-off-by: Rik van Riel <riel@redhat.com> ---- This is against tip.git numa/core diff --git a/CREDITS b/CREDITS index d8fe12a..b4cdc8f 100644 --- a/CREDITS +++ b/CREDITS @@ -125,6 +125,7 @@ D: Author of pscan that helps to fix lp/parport bugs D: Author of lil (Linux Interrupt Latency benchmark) D: Fixed the shm swap deallocation at swapoff time (try_to_unuse message) D: VM hacker +D: NUMA task placement D: Various other kernel hacks S: Imola 40026 S: Italy diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 1e24aa1..e93032d 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -18,6 +18,9 @@ * * Adaptive scheduling granularity, math enhancements by Peter Zijlstra * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * + * NUMA placement, statistics and algorithm by Andrea Arcangeli, + * CFS balancing changes by Peter Zijlstra. Copyright (C) 2012 Red Hat, Inc. */ #include <linux/latencytop.h> diff --git a/mm/memory.c b/mm/memory.c index fc48fe8..9e56a44 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -36,6 +36,8 @@ * (Gerhard.Wichert@pdb.siemens.de) * * Aug/Sep 2004 Changed to four level page tables (Andi Kleen) + * + * 2012 - NUMA placement page faults (Andrea Arcangeli, Peter Zijlstra) */ #include <linux/kernel_stat.h> ^ permalink raw reply related [flat|nested] 20+ messages in thread
* Re: [PATCH 1/2] add credits for NUMA placement 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel @ 2012-10-19 11:41 ` Peter Zijlstra 2012-10-19 12:02 ` [tip:numa/core] numa: Add " tip-bot for Rik van Riel ` (4 subsequent siblings) 5 siblings, 0 replies; 20+ messages in thread From: Peter Zijlstra @ 2012-10-19 11:41 UTC (permalink / raw) To: Rik van Riel; +Cc: mingo, linux-kernel, aarcange On Thu, 2012-10-18 at 17:19 -0400, Rik van Riel wrote: > The NUMA placement code has been rewritten several times, but > the basic ideas took a lot of work to develop. The people who > put in the work deserve credit for it. Thanks Andrea & Peter :) > > The Documentation/scheduler/numa-problem.txt file should > probably be rewritten once we figure out the final details of > what the NUMA code needs to do, and why. > > Signed-off-by: Rik van Riel <riel@redhat.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Thanks Rik! ^ permalink raw reply [flat|nested] 20+ messages in thread
* [tip:numa/core] numa: Add credits for NUMA placement 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel 2012-10-19 11:41 ` Peter Zijlstra @ 2012-10-19 12:02 ` tip-bot for Rik van Riel 2012-10-28 17:11 ` [tip:numa/core] sched, numa, mm: " tip-bot for Rik van Riel ` (3 subsequent siblings) 5 siblings, 0 replies; 20+ messages in thread From: tip-bot for Rik van Riel @ 2012-10-19 12:02 UTC (permalink / raw) To: linux-tip-commits Cc: linux-kernel, hpa, mingo, torvalds, a.p.zijlstra, riel, akpm, tglx Commit-ID: c1a305006e4dd428001852923c11806d754db9f1 Gitweb: http://git.kernel.org/tip/c1a305006e4dd428001852923c11806d754db9f1 Author: Rik van Riel <riel@redhat.com> AuthorDate: Thu, 18 Oct 2012 17:19:28 -0400 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Fri, 19 Oct 2012 13:45:48 +0200 numa: Add credits for NUMA placement The NUMA placement code has been rewritten several times, but the basic ideas took a lot of work to develop. The people who put in the work deserve credit for it. Thanks Andrea & Peter :) [ The Documentation/scheduler/numa-problem.txt file should probably be rewritten once we figure out the final details of what the NUMA code needs to do, and why. ] Signed-off-by: Rik van Riel <riel@redhat.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: aarcange@redhat.com Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Link: http://lkml.kernel.org/r/20121018171928.24d06af4@cuia.bos.redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org> ---- This is against tip.git numa/core --- CREDITS | 1 + kernel/sched/fair.c | 3 +++ mm/memory.c | 2 ++ 3 files changed, 6 insertions(+), 0 deletions(-) diff --git a/CREDITS b/CREDITS index d8fe12a..b4cdc8f 100644 --- a/CREDITS +++ b/CREDITS @@ -125,6 +125,7 @@ D: Author of pscan that helps to fix lp/parport bugs D: Author of lil (Linux Interrupt Latency benchmark) D: Fixed the shm swap deallocation at swapoff time (try_to_unuse message) D: VM hacker +D: NUMA task placement D: Various other kernel hacks S: Imola 40026 S: Italy diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 1e24aa1..e93032d 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -18,6 +18,9 @@ * * Adaptive scheduling granularity, math enhancements by Peter Zijlstra * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * + * NUMA placement, statistics and algorithm by Andrea Arcangeli, + * CFS balancing changes by Peter Zijlstra. Copyright (C) 2012 Red Hat, Inc. */ #include <linux/latencytop.h> diff --git a/mm/memory.c b/mm/memory.c index fc48fe8..9e56a44 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -36,6 +36,8 @@ * (Gerhard.Wichert@pdb.siemens.de) * * Aug/Sep 2004 Changed to four level page tables (Andi Kleen) + * + * 2012 - NUMA placement page faults (Andrea Arcangeli, Peter Zijlstra) */ #include <linux/kernel_stat.h> ^ permalink raw reply related [flat|nested] 20+ messages in thread
* [tip:numa/core] sched, numa, mm: Add credits for NUMA placement 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel 2012-10-19 11:41 ` Peter Zijlstra 2012-10-19 12:02 ` [tip:numa/core] numa: Add " tip-bot for Rik van Riel @ 2012-10-28 17:11 ` tip-bot for Rik van Riel 2012-11-13 15:31 ` tip-bot for Rik van Riel ` (2 subsequent siblings) 5 siblings, 0 replies; 20+ messages in thread From: tip-bot for Rik van Riel @ 2012-10-28 17:11 UTC (permalink / raw) To: linux-tip-commits Cc: linux-kernel, hpa, mingo, torvalds, a.p.zijlstra, riel, akpm, tglx Commit-ID: c2ef354e5ab9d06a6b914f1241ade0b681330ffb Gitweb: http://git.kernel.org/tip/c2ef354e5ab9d06a6b914f1241ade0b681330ffb Author: Rik van Riel <riel@redhat.com> AuthorDate: Thu, 18 Oct 2012 17:19:28 -0400 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Sun, 28 Oct 2012 17:31:16 +0100 sched, numa, mm: Add credits for NUMA placement The NUMA placement code has been rewritten several times, but the basic ideas took a lot of work to develop. The people who put in the work deserve credit for it. Thanks Andrea & Peter :) [ The Documentation/scheduler/numa-problem.txt file should probably be rewritten once we figure out the final details of what the NUMA code needs to do, and why. ] Signed-off-by: Rik van Riel <riel@redhat.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: aarcange@redhat.com Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Link: http://lkml.kernel.org/r/20121018171928.24d06af4@cuia.bos.redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org> ---- This is against tip.git numa/core --- CREDITS | 1 + kernel/sched/fair.c | 3 +++ mm/memory.c | 2 ++ 3 files changed, 6 insertions(+), 0 deletions(-) diff --git a/CREDITS b/CREDITS index d8fe12a..b4cdc8f 100644 --- a/CREDITS +++ b/CREDITS @@ -125,6 +125,7 @@ D: Author of pscan that helps to fix lp/parport bugs D: Author of lil (Linux Interrupt Latency benchmark) D: Fixed the shm swap deallocation at swapoff time (try_to_unuse message) D: VM hacker +D: NUMA task placement D: Various other kernel hacks S: Imola 40026 S: Italy diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index bfe07cd..3e51cfd 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -18,6 +18,9 @@ * * Adaptive scheduling granularity, math enhancements by Peter Zijlstra * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * + * NUMA placement, statistics and algorithm by Andrea Arcangeli, + * CFS balancing changes by Peter Zijlstra. Copyright (C) 2012 Red Hat, Inc. */ #include <linux/latencytop.h> diff --git a/mm/memory.c b/mm/memory.c index 3ecfeca..2c17d82 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -36,6 +36,8 @@ * (Gerhard.Wichert@pdb.siemens.de) * * Aug/Sep 2004 Changed to four level page tables (Andi Kleen) + * + * 2012 - NUMA placement page faults (Andrea Arcangeli, Peter Zijlstra) */ #include <linux/kernel_stat.h> ^ permalink raw reply related [flat|nested] 20+ messages in thread
* [tip:numa/core] sched, numa, mm: Add credits for NUMA placement 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel ` (2 preceding siblings ...) 2012-10-28 17:11 ` [tip:numa/core] sched, numa, mm: " tip-bot for Rik van Riel @ 2012-11-13 15:31 ` tip-bot for Rik van Riel 2012-11-13 17:23 ` tip-bot for Rik van Riel 2012-11-19 19:46 ` tip-bot for Rik van Riel 5 siblings, 0 replies; 20+ messages in thread From: tip-bot for Rik van Riel @ 2012-11-13 15:31 UTC (permalink / raw) To: linux-tip-commits Cc: linux-kernel, hpa, mingo, torvalds, a.p.zijlstra, riel, akpm, aarcange, mgorman, tglx Commit-ID: 0c4a966d7968363c833e58068da58a121095b075 Gitweb: http://git.kernel.org/tip/0c4a966d7968363c833e58068da58a121095b075 Author: Rik van Riel <riel@redhat.com> AuthorDate: Thu, 18 Oct 2012 17:19:28 -0400 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Tue, 13 Nov 2012 14:11:50 +0100 sched, numa, mm: Add credits for NUMA placement The NUMA placement code has been rewritten several times, but the basic ideas took a lot of work to develop. The people who put in the work deserve credit for it. Thanks Andrea & Peter :) [ The Documentation/scheduler/numa-problem.txt file should probably be rewritten once we figure out the final details of what the NUMA code needs to do, and why. ] Signed-off-by: Rik van Riel <riel@redhat.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Link: http://lkml.kernel.org/r/20121018171928.24d06af4@cuia.bos.redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org> ---- This is against tip.git numa/core --- CREDITS | 1 + kernel/sched/fair.c | 3 +++ mm/memory.c | 2 ++ 3 files changed, 6 insertions(+), 0 deletions(-) diff --git a/CREDITS b/CREDITS index d8fe12a..b4cdc8f 100644 --- a/CREDITS +++ b/CREDITS @@ -125,6 +125,7 @@ D: Author of pscan that helps to fix lp/parport bugs D: Author of lil (Linux Interrupt Latency benchmark) D: Fixed the shm swap deallocation at swapoff time (try_to_unuse message) D: VM hacker +D: NUMA task placement D: Various other kernel hacks S: Imola 40026 S: Italy diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 93f4de4..309a254 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -18,6 +18,9 @@ * * Adaptive scheduling granularity, math enhancements by Peter Zijlstra * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * + * NUMA placement, statistics and algorithm by Andrea Arcangeli, + * CFS balancing changes by Peter Zijlstra. Copyright (C) 2012 Red Hat, Inc. */ #include <linux/latencytop.h> diff --git a/mm/memory.c b/mm/memory.c index 1b9108c..ebd18fd 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -36,6 +36,8 @@ * (Gerhard.Wichert@pdb.siemens.de) * * Aug/Sep 2004 Changed to four level page tables (Andi Kleen) + * + * 2012 - NUMA placement page faults (Andrea Arcangeli, Peter Zijlstra) */ #include <linux/kernel_stat.h> ^ permalink raw reply related [flat|nested] 20+ messages in thread
* [tip:numa/core] sched, numa, mm: Add credits for NUMA placement 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel ` (3 preceding siblings ...) 2012-11-13 15:31 ` tip-bot for Rik van Riel @ 2012-11-13 17:23 ` tip-bot for Rik van Riel 2012-11-19 19:46 ` tip-bot for Rik van Riel 5 siblings, 0 replies; 20+ messages in thread From: tip-bot for Rik van Riel @ 2012-11-13 17:23 UTC (permalink / raw) To: linux-tip-commits Cc: linux-kernel, hpa, mingo, torvalds, a.p.zijlstra, riel, akpm, aarcange, mgorman, tglx Commit-ID: b644797f6b72b9b9b0cf35bdf7981f5602725bea Gitweb: http://git.kernel.org/tip/b644797f6b72b9b9b0cf35bdf7981f5602725bea Author: Rik van Riel <riel@redhat.com> AuthorDate: Thu, 18 Oct 2012 17:19:28 -0400 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Tue, 13 Nov 2012 18:09:25 +0100 sched, numa, mm: Add credits for NUMA placement The NUMA placement code has been rewritten several times, but the basic ideas took a lot of work to develop. The people who put in the work deserve credit for it. Thanks Andrea & Peter :) [ The Documentation/scheduler/numa-problem.txt file should probably be rewritten once we figure out the final details of what the NUMA code needs to do, and why. ] Signed-off-by: Rik van Riel <riel@redhat.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Link: http://lkml.kernel.org/r/20121018171928.24d06af4@cuia.bos.redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org> ---- This is against tip.git numa/core --- CREDITS | 1 + kernel/sched/fair.c | 3 +++ mm/memory.c | 2 ++ 3 files changed, 6 insertions(+), 0 deletions(-) diff --git a/CREDITS b/CREDITS index d8fe12a..b4cdc8f 100644 --- a/CREDITS +++ b/CREDITS @@ -125,6 +125,7 @@ D: Author of pscan that helps to fix lp/parport bugs D: Author of lil (Linux Interrupt Latency benchmark) D: Fixed the shm swap deallocation at swapoff time (try_to_unuse message) D: VM hacker +D: NUMA task placement D: Various other kernel hacks S: Imola 40026 S: Italy diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 93f4de4..309a254 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -18,6 +18,9 @@ * * Adaptive scheduling granularity, math enhancements by Peter Zijlstra * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * + * NUMA placement, statistics and algorithm by Andrea Arcangeli, + * CFS balancing changes by Peter Zijlstra. Copyright (C) 2012 Red Hat, Inc. */ #include <linux/latencytop.h> diff --git a/mm/memory.c b/mm/memory.c index 1b9108c..ebd18fd 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -36,6 +36,8 @@ * (Gerhard.Wichert@pdb.siemens.de) * * Aug/Sep 2004 Changed to four level page tables (Andi Kleen) + * + * 2012 - NUMA placement page faults (Andrea Arcangeli, Peter Zijlstra) */ #include <linux/kernel_stat.h> ^ permalink raw reply related [flat|nested] 20+ messages in thread
* [tip:numa/core] sched, numa, mm: Add credits for NUMA placement 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel ` (4 preceding siblings ...) 2012-11-13 17:23 ` tip-bot for Rik van Riel @ 2012-11-19 19:46 ` tip-bot for Rik van Riel 5 siblings, 0 replies; 20+ messages in thread From: tip-bot for Rik van Riel @ 2012-11-19 19:46 UTC (permalink / raw) To: linux-tip-commits Cc: linux-kernel, hpa, mingo, torvalds, a.p.zijlstra, hughd, riel, akpm, aarcange, mgorman, tglx Commit-ID: cb27d6087bc812d0624ef774a9ddee81f7cc0895 Gitweb: http://git.kernel.org/tip/cb27d6087bc812d0624ef774a9ddee81f7cc0895 Author: Rik van Riel <riel@redhat.com> AuthorDate: Thu, 18 Oct 2012 17:19:28 -0400 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Mon, 19 Nov 2012 03:31:54 +0100 sched, numa, mm: Add credits for NUMA placement The NUMA placement code has been rewritten several times, but the basic ideas took a lot of work to develop. The people who put in the work deserve credit for it. Thanks Andrea & Peter :) [ The Documentation/scheduler/numa-problem.txt file should probably be rewritten once we figure out the final details of what the NUMA code needs to do, and why. ] Signed-off-by: Rik van Riel <riel@redhat.com> Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20121018171928.24d06af4@cuia.bos.redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org> ---- This is against tip.git numa/core --- CREDITS | 1 + kernel/sched/fair.c | 3 +++ mm/memory.c | 2 ++ 3 files changed, 6 insertions(+) diff --git a/CREDITS b/CREDITS index d8fe12a..b4cdc8f 100644 --- a/CREDITS +++ b/CREDITS @@ -125,6 +125,7 @@ D: Author of pscan that helps to fix lp/parport bugs D: Author of lil (Linux Interrupt Latency benchmark) D: Fixed the shm swap deallocation at swapoff time (try_to_unuse message) D: VM hacker +D: NUMA task placement D: Various other kernel hacks S: Imola 40026 S: Italy diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 511fbb8..8af0208 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -18,6 +18,9 @@ * * Adaptive scheduling granularity, math enhancements by Peter Zijlstra * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> + * + * NUMA placement, statistics and algorithm by Andrea Arcangeli, + * CFS balancing changes by Peter Zijlstra. Copyright (C) 2012 Red Hat, Inc. */ #include <linux/latencytop.h> diff --git a/mm/memory.c b/mm/memory.c index 52ad29d..1f733dc 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -36,6 +36,8 @@ * (Gerhard.Wichert@pdb.siemens.de) * * Aug/Sep 2004 Changed to four level page tables (Andi Kleen) + * + * 2012 - NUMA placement page faults (Andrea Arcangeli, Peter Zijlstra) */ #include <linux/kernel_stat.h> ^ permalink raw reply related [flat|nested] 20+ messages in thread
* [PATCH 2/2] rename NUMA fault handling functions 2012-10-18 21:18 [PATCH 0/2] minor NUMA cleanups & documentation Rik van Riel 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel @ 2012-10-18 21:20 ` Rik van Riel 2012-10-19 11:41 ` Peter Zijlstra ` (2 more replies) 2012-10-21 12:43 ` [PATCH 3/2] sched, numa, mm: Implement constant rate working set sampling Ingo Molnar 2 siblings, 3 replies; 20+ messages in thread From: Rik van Riel @ 2012-10-18 21:20 UTC (permalink / raw) To: mingo; +Cc: linux-kernel, aarcange, a.p.zijlstra Having the function name indicate what the function is used for makes the code a little easier to read. Furthermore, the fault handling code largely consists of do_...._page functions. Rename the NUMA fault handling functions to indicate what they are used for. Signed-off-by: Rik van Riel <riel@redhat.com> --- Against tip.git numa/core include/linux/huge_mm.h | 8 ++++---- mm/huge_memory.c | 4 ++-- mm/memory.c | 18 ++++++++++-------- 3 files changed, 16 insertions(+), 14 deletions(-) diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index ed60d79..9580e22 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -161,9 +161,9 @@ static inline struct page *compound_trans_head(struct page *page) return page; } -extern bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd); +extern bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd); -extern void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +extern void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t orig_pmd); @@ -204,12 +204,12 @@ static inline int pmd_trans_huge_lock(pmd_t *pmd, return 0; } -static inline bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) +static inline bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd) { return false; } -static inline void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +static inline void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t orig_pmd) { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 5afd0d7..c25fd37 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -751,7 +751,7 @@ out: return handle_pte_fault(mm, vma, address, pte, pmd, flags); } -bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) +bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd) { /* * See pte_prot_none(). @@ -762,7 +762,7 @@ bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) return pmd_same(pmd, pmd_modify(pmd, vma_prot_none(vma))); } -void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t entry) { diff --git a/mm/memory.c b/mm/memory.c index 9e56a44..c752379 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -3425,11 +3425,13 @@ static int do_nonlinear_fault(struct mm_struct *mm, struct vm_area_struct *vma, return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte); } -static bool pte_prot_none(struct vm_area_struct *vma, pte_t pte) +static bool pte_numa(struct vm_area_struct *vma, pte_t pte) { /* - * If we have the normal vma->vm_page_prot protections we're not a - * 'special' PROT_NONE page. + * For NUMA page faults, we use PROT_NONE ptes in VMAs with + * "normal" vma->vm_page_prot protections. Genuine PROT_NONE + * VMAs should never get here, because the fault handling code + * will notice that the VMA has no read or write permissions. * * This means we cannot get 'special' PROT_NONE faults from genuine * PROT_NONE maps, nor from PROT_WRITE file maps that do dirty @@ -3444,7 +3446,7 @@ static bool pte_prot_none(struct vm_area_struct *vma, pte_t pte) return pte_same(pte, pte_modify(pte, vma_prot_none(vma))); } -static int do_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +static int do_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *ptep, pmd_t *pmd, unsigned int flags, pte_t entry) { @@ -3541,8 +3543,8 @@ int handle_pte_fault(struct mm_struct *mm, pte, pmd, flags, entry); } - if (pte_prot_none(vma, entry)) - return do_prot_none(mm, vma, address, pte, pmd, flags, entry); + if (pte_numa(vma, entry)) + return do_numa_page(mm, vma, address, pte, pmd, flags, entry); ptl = pte_lockptr(mm, pmd); spin_lock(ptl); @@ -3612,8 +3614,8 @@ retry: barrier(); if (pmd_trans_huge(orig_pmd) && !pmd_trans_splitting(orig_pmd)) { - if (pmd_prot_none(vma, orig_pmd)) { - do_huge_pmd_prot_none(mm, vma, address, pmd, + if (pmd_numa(vma, orig_pmd)) { + do_huge_pmd_numa_page(mm, vma, address, pmd, flags, orig_pmd); } ^ permalink raw reply related [flat|nested] 20+ messages in thread
* Re: [PATCH 2/2] rename NUMA fault handling functions 2012-10-18 21:20 ` [PATCH 2/2] rename NUMA fault handling functions Rik van Riel @ 2012-10-19 11:41 ` Peter Zijlstra 2012-10-19 14:07 ` Rik van Riel 2012-10-20 10:15 ` [PATCH 2/2] rename NUMA " Michel Lespinasse 2012-10-21 15:20 ` [tip:numa/core] numa, mm: Rename the PROT_NONE fault handling functions to *_numa() tip-bot for Rik van Riel 2012-10-23 11:00 ` tip-bot for Rik van Riel 2 siblings, 2 replies; 20+ messages in thread From: Peter Zijlstra @ 2012-10-19 11:41 UTC (permalink / raw) To: Rik van Riel; +Cc: mingo, linux-kernel, aarcange, Michel Lespinasse On Thu, 2012-10-18 at 17:20 -0400, Rik van Riel wrote: > Having the function name indicate what the function is used > for makes the code a little easier to read. Furthermore, > the fault handling code largely consists of do_...._page > functions. I don't much care either way, but I was thinking walken might want to use something similar to do WSS estimation, in which case the NUMA name is just as wrong. ^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [PATCH 2/2] rename NUMA fault handling functions 2012-10-19 11:41 ` Peter Zijlstra @ 2012-10-19 14:07 ` Rik van Riel 2012-10-19 20:54 ` Ingo Molnar 2012-10-20 10:15 ` [PATCH 2/2] rename NUMA " Michel Lespinasse 1 sibling, 1 reply; 20+ messages in thread From: Rik van Riel @ 2012-10-19 14:07 UTC (permalink / raw) To: Peter Zijlstra; +Cc: mingo, linux-kernel, aarcange, Michel Lespinasse On 10/19/2012 07:41 AM, Peter Zijlstra wrote: > On Thu, 2012-10-18 at 17:20 -0400, Rik van Riel wrote: >> Having the function name indicate what the function is used >> for makes the code a little easier to read. Furthermore, >> the fault handling code largely consists of do_...._page >> functions. > > I don't much care either way, but I was thinking walken might want to > use something similar to do WSS estimation, in which case the NUMA name > is just as wrong. That's a good point. I had not considered other uses of the same code. -- All rights reversed ^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [PATCH 2/2] rename NUMA fault handling functions 2012-10-19 14:07 ` Rik van Riel @ 2012-10-19 20:54 ` Ingo Molnar 2012-10-21 12:50 ` [PATCH 4/2] numa, mm: Rename the PROT_NONE " Ingo Molnar 0 siblings, 1 reply; 20+ messages in thread From: Ingo Molnar @ 2012-10-19 20:54 UTC (permalink / raw) To: Rik van Riel Cc: Peter Zijlstra, mingo, linux-kernel, aarcange, Michel Lespinasse * Rik van Riel <riel@redhat.com> wrote: > On 10/19/2012 07:41 AM, Peter Zijlstra wrote: > >On Thu, 2012-10-18 at 17:20 -0400, Rik van Riel wrote: > >>Having the function name indicate what the function is used > >>for makes the code a little easier to read. Furthermore, > >>the fault handling code largely consists of do_...._page > >>functions. > > > > I don't much care either way, but I was thinking walken > > might want to use something similar to do WSS estimation, in > > which case the NUMA name is just as wrong. > > That's a good point. I had not considered other uses of the > same code. Renaming the functions for more clarity still makes sense IMO: we could give it a _wss or _working_set prefix/postfix? Thanks, Ingo ^ permalink raw reply [flat|nested] 20+ messages in thread
* [PATCH 4/2] numa, mm: Rename the PROT_NONE fault handling functions 2012-10-19 20:54 ` Ingo Molnar @ 2012-10-21 12:50 ` Ingo Molnar 2012-10-21 13:22 ` Rik van Riel 0 siblings, 1 reply; 20+ messages in thread From: Ingo Molnar @ 2012-10-21 12:50 UTC (permalink / raw) To: Rik van Riel Cc: Peter Zijlstra, mingo, linux-kernel, aarcange, Michel Lespinasse * Ingo Molnar <mingo@kernel.org> wrote: > > > I don't much care either way, but I was thinking walken > > > might want to use something similar to do WSS estimation, > > > in which case the NUMA name is just as wrong. > > > > That's a good point. I had not considered other uses of the > > same code. > > Renaming the functions for more clarity still makes sense IMO: > we could give it a _wss or _working_set prefix/postfix? So, to not drop your patch on the floor I've modified it as per the patch below. The _wss() names signal that these handlers are used for a specific purpose, they are not related to the regular PROT_NONE handling code. Agreed? Thanks, Ingo ---------------> >From 7e426e0f6ffe228118e57a70ae402e21792a0456 Mon Sep 17 00:00:00 2001 From: Rik van Riel <riel@redhat.com> Date: Thu, 18 Oct 2012 17:20:21 -0400 Subject: [PATCH] numa, mm: Rename the PROT_NONE fault handling functions Having the function name indicate what the function is used for makes the code a little easier to read. Furthermore, the fault handling code largely consists of do_...._page functions. Rename the Working-Set Sampling (WSS) fault handling functions to indicate what they are used for. Signed-off-by: Rik van Riel <riel@redhat.com> Cc: aarcange@redhat.com Cc: a.p.zijlstra@chello.nl Link: http://lkml.kernel.org/r/20121018172021.0b1f6e3d@cuia.bos.redhat.com [ Changed the naming pattern to 'working-set sampling (WSS)' wss_() ] Signed-off-by: Ingo Molnar <mingo@kernel.org> --- include/linux/huge_mm.h | 8 ++++---- mm/huge_memory.c | 4 ++-- mm/memory.c | 18 ++++++++++-------- 3 files changed, 16 insertions(+), 14 deletions(-) diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index bcbe467..93c6ab5 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -160,9 +160,9 @@ static inline struct page *compound_trans_head(struct page *page) return page; } -extern bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd); +extern bool pmd_wss(struct vm_area_struct *vma, pmd_t pmd); -extern void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +extern void do_huge_pmd_wss_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t orig_pmd); @@ -203,12 +203,12 @@ static inline int pmd_trans_huge_lock(pmd_t *pmd, return 0; } -static inline bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) +static inline bool pmd_wss(struct vm_area_struct *vma, pmd_t pmd) { return false; } -static inline void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +static inline void do_huge_pmd_wss_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t orig_pmd) { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index c58a5f0..982f678 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -727,7 +727,7 @@ out: return handle_pte_fault(mm, vma, address, pte, pmd, flags); } -bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) +bool pmd_wss(struct vm_area_struct *vma, pmd_t pmd) { /* * See pte_prot_none(). @@ -738,7 +738,7 @@ bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) return pmd_same(pmd, pmd_modify(pmd, vma_prot_none(vma))); } -void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +void do_huge_pmd_wss_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t entry) { diff --git a/mm/memory.c b/mm/memory.c index 2cc8a29..a3693e6 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1471,11 +1471,13 @@ int zap_vma_ptes(struct vm_area_struct *vma, unsigned long address, } EXPORT_SYMBOL_GPL(zap_vma_ptes); -static bool pte_prot_none(struct vm_area_struct *vma, pte_t pte) +static bool pte_wss(struct vm_area_struct *vma, pte_t pte) { /* - * If we have the normal vma->vm_page_prot protections we're not a - * 'special' PROT_NONE page. + * For NUMA page faults, we use PROT_NONE ptes in VMAs with + * "normal" vma->vm_page_prot protections. Genuine PROT_NONE + * VMAs should never get here, because the fault handling code + * will notice that the VMA has no read or write permissions. * * This means we cannot get 'special' PROT_NONE faults from genuine * PROT_NONE maps, nor from PROT_WRITE file maps that do dirty @@ -3476,7 +3478,7 @@ static int do_nonlinear_fault(struct mm_struct *mm, struct vm_area_struct *vma, return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte); } -static int do_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +static int do_wss_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *ptep, pmd_t *pmd, unsigned int flags, pte_t entry) { @@ -3601,8 +3603,8 @@ int handle_pte_fault(struct mm_struct *mm, pte, pmd, flags, entry); } - if (pte_prot_none(vma, entry)) - return do_prot_none(mm, vma, address, pte, pmd, flags, entry); + if (pte_wss(vma, entry)) + return do_wss_page(mm, vma, address, pte, pmd, flags, entry); ptl = pte_lockptr(mm, pmd); spin_lock(ptl); @@ -3672,8 +3674,8 @@ retry: barrier(); if (pmd_trans_huge(orig_pmd) && !pmd_trans_splitting(orig_pmd)) { - if (pmd_prot_none(vma, orig_pmd)) { - do_huge_pmd_prot_none(mm, vma, address, pmd, + if (pmd_wss(vma, orig_pmd)) { + do_huge_pmd_wss_page(mm, vma, address, pmd, flags, orig_pmd); } ^ permalink raw reply related [flat|nested] 20+ messages in thread
* Re: [PATCH 4/2] numa, mm: Rename the PROT_NONE fault handling functions 2012-10-21 12:50 ` [PATCH 4/2] numa, mm: Rename the PROT_NONE " Ingo Molnar @ 2012-10-21 13:22 ` Rik van Riel 2012-10-21 13:29 ` Ingo Molnar 0 siblings, 1 reply; 20+ messages in thread From: Rik van Riel @ 2012-10-21 13:22 UTC (permalink / raw) To: Ingo Molnar Cc: Peter Zijlstra, mingo, linux-kernel, aarcange, Michel Lespinasse On 10/21/2012 08:50 AM, Ingo Molnar wrote: > > * Ingo Molnar <mingo@kernel.org> wrote: > >>>> I don't much care either way, but I was thinking walken >>>> might want to use something similar to do WSS estimation, >>>> in which case the NUMA name is just as wrong. >>> >>> That's a good point. I had not considered other uses of the >>> same code. >> >> Renaming the functions for more clarity still makes sense IMO: >> we could give it a _wss or _working_set prefix/postfix? > > So, to not drop your patch on the floor I've modified it as per > the patch below. > > The _wss() names signal that these handlers are used for a > specific purpose, they are not related to the regular PROT_NONE > handling code. Michel indicated that he does not use PROT_NONE for his working set estimation code, but instead checks the accessed bits in the page tables. Since NUMA migration is the only user of PROT_NONE ptes in normal vmas, maybe _numa is the right suffix after all? -- All rights reversed ^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [PATCH 4/2] numa, mm: Rename the PROT_NONE fault handling functions 2012-10-21 13:22 ` Rik van Riel @ 2012-10-21 13:29 ` Ingo Molnar 2012-10-21 13:43 ` Ingo Molnar 0 siblings, 1 reply; 20+ messages in thread From: Ingo Molnar @ 2012-10-21 13:29 UTC (permalink / raw) To: Rik van Riel Cc: Peter Zijlstra, mingo, linux-kernel, aarcange, Michel Lespinasse * Rik van Riel <riel@redhat.com> wrote: > On 10/21/2012 08:50 AM, Ingo Molnar wrote: > > > >* Ingo Molnar <mingo@kernel.org> wrote: > > > >>>>I don't much care either way, but I was thinking walken > >>>>might want to use something similar to do WSS estimation, > >>>>in which case the NUMA name is just as wrong. > >>> > >>>That's a good point. I had not considered other uses of the > >>>same code. > >> > >>Renaming the functions for more clarity still makes sense IMO: > >>we could give it a _wss or _working_set prefix/postfix? > > > >So, to not drop your patch on the floor I've modified it as per > >the patch below. > > > >The _wss() names signal that these handlers are used for a > >specific purpose, they are not related to the regular PROT_NONE > >handling code. > > Michel indicated that he does not use PROT_NONE for his > working set estimation code, but instead checks the accessed > bits in the page tables. The pte_young() WSS method has a couple of fundamental limitations: - it doesn't work with shared memory very well, the pte is per mapping, not per page. The PROT_NONE method instruments the physical page in essence. - it does not tell us which task touched the pte, in a multi-threaded program So like Peter I'd too expect these new WSS methods to eventually be picked up for any serious WSS work. Thanks, Ingo ^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [PATCH 4/2] numa, mm: Rename the PROT_NONE fault handling functions 2012-10-21 13:29 ` Ingo Molnar @ 2012-10-21 13:43 ` Ingo Molnar 0 siblings, 0 replies; 20+ messages in thread From: Ingo Molnar @ 2012-10-21 13:43 UTC (permalink / raw) To: Rik van Riel Cc: Peter Zijlstra, mingo, linux-kernel, aarcange, Michel Lespinasse * Ingo Molnar <mingo@kernel.org> wrote: > > Michel indicated that he does not use PROT_NONE for his > > working set estimation code, but instead checks the accessed > > bits in the page tables. > > The pte_young() WSS method has a couple of fundamental > limitations: > > - it doesn't work with shared memory very well, the pte is per > mapping, not per page. The PROT_NONE method instruments the > physical page in essence. > > - it does not tell us which task touched the pte, in a > multi-threaded program > > So like Peter I'd too expect these new WSS methods to eventually > be picked up for any serious WSS work. Nevertheless lets wait and see until it actually happens - and meanwhile the prot_none namings are confusing. So I've applied your patch as-is, with two more (new) usage sites converted as well. Will push it out after a bit of testing. Thanks, Ingo ^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [PATCH 2/2] rename NUMA fault handling functions 2012-10-19 11:41 ` Peter Zijlstra 2012-10-19 14:07 ` Rik van Riel @ 2012-10-20 10:15 ` Michel Lespinasse 1 sibling, 0 replies; 20+ messages in thread From: Michel Lespinasse @ 2012-10-20 10:15 UTC (permalink / raw) To: Peter Zijlstra; +Cc: Rik van Riel, mingo, linux-kernel, aarcange On Fri, Oct 19, 2012 at 4:41 AM, Peter Zijlstra <a.p.zijlstra@chello.nl> wrote: > On Thu, 2012-10-18 at 17:20 -0400, Rik van Riel wrote: >> Having the function name indicate what the function is used >> for makes the code a little easier to read. Furthermore, >> the fault handling code largely consists of do_...._page >> functions. > > I don't much care either way, but I was thinking walken might want to > use something similar to do WSS estimation, in which case the NUMA name > is just as wrong. Right now my working set estimation only uses A bits, so let's not make that a concern here. I think the _numa names are a bit better than _prot_none, but still a bit confusing. I don't have any great suggestions but I think there should at least be a comment above pte_numa() that explains what the numa ptes are (the comment within the function doesn't qualify as it only explains how the numa ptes are different from the ones in PROT_NONE vmas...) -- Michel "Walken" Lespinasse A program is never fully debugged until the last user dies. ^ permalink raw reply [flat|nested] 20+ messages in thread
* [tip:numa/core] numa, mm: Rename the PROT_NONE fault handling functions to *_numa() 2012-10-18 21:20 ` [PATCH 2/2] rename NUMA fault handling functions Rik van Riel 2012-10-19 11:41 ` Peter Zijlstra @ 2012-10-21 15:20 ` tip-bot for Rik van Riel 2012-10-23 11:00 ` tip-bot for Rik van Riel 2 siblings, 0 replies; 20+ messages in thread From: tip-bot for Rik van Riel @ 2012-10-21 15:20 UTC (permalink / raw) To: linux-tip-commits; +Cc: linux-kernel, riel, hpa, mingo, tglx Commit-ID: 2458840fddea542391d343dac734d149607db709 Gitweb: http://git.kernel.org/tip/2458840fddea542391d343dac734d149607db709 Author: Rik van Riel <riel@redhat.com> AuthorDate: Thu, 18 Oct 2012 17:20:21 -0400 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Sun, 21 Oct 2012 15:41:26 +0200 numa, mm: Rename the PROT_NONE fault handling functions to *_numa() Having the function name indicate what the function is used for makes the code a little easier to read. Furthermore, the fault handling code largely consists of do_...._page functions. Rename the NUMA working set sampling fault handling functions to _numa() names, to indicate what they are used for. This separates the naming from the regular PROT_NONE namings. Signed-off-by: Rik van Riel <riel@redhat.com> Cc: aarcange@redhat.com Cc: a.p.zijlstra@chello.nl Link: http://lkml.kernel.org/r/20121018172021.0b1f6e3d@cuia.bos.redhat.com [ Converted two more usage sites ] Signed-off-by: Ingo Molnar <mingo@kernel.org> --- include/linux/huge_mm.h | 8 ++++---- mm/huge_memory.c | 4 ++-- mm/memory.c | 22 ++++++++++++---------- 3 files changed, 18 insertions(+), 16 deletions(-) diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index bcbe467..4f0f948 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -160,9 +160,9 @@ static inline struct page *compound_trans_head(struct page *page) return page; } -extern bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd); +extern bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd); -extern void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +extern void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t orig_pmd); @@ -203,12 +203,12 @@ static inline int pmd_trans_huge_lock(pmd_t *pmd, return 0; } -static inline bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) +static inline bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd) { return false; } -static inline void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +static inline void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t orig_pmd) { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index c58a5f0..a8f6531 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -727,7 +727,7 @@ out: return handle_pte_fault(mm, vma, address, pte, pmd, flags); } -bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) +bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd) { /* * See pte_prot_none(). @@ -738,7 +738,7 @@ bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) return pmd_same(pmd, pmd_modify(pmd, vma_prot_none(vma))); } -void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t entry) { diff --git a/mm/memory.c b/mm/memory.c index 2cc8a29..23d4bd4 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1471,11 +1471,13 @@ int zap_vma_ptes(struct vm_area_struct *vma, unsigned long address, } EXPORT_SYMBOL_GPL(zap_vma_ptes); -static bool pte_prot_none(struct vm_area_struct *vma, pte_t pte) +static bool pte_numa(struct vm_area_struct *vma, pte_t pte) { /* - * If we have the normal vma->vm_page_prot protections we're not a - * 'special' PROT_NONE page. + * For NUMA page faults, we use PROT_NONE ptes in VMAs with + * "normal" vma->vm_page_prot protections. Genuine PROT_NONE + * VMAs should never get here, because the fault handling code + * will notice that the VMA has no read or write permissions. * * This means we cannot get 'special' PROT_NONE faults from genuine * PROT_NONE maps, nor from PROT_WRITE file maps that do dirty @@ -1543,7 +1545,7 @@ struct page *follow_page(struct vm_area_struct *vma, unsigned long address, page = follow_huge_pmd(mm, address, pmd, flags & FOLL_WRITE); goto out; } - if ((flags & FOLL_NUMA) && pmd_prot_none(vma, *pmd)) + if ((flags & FOLL_NUMA) && pmd_numa(vma, *pmd)) goto no_page_table; if (pmd_trans_huge(*pmd)) { if (flags & FOLL_SPLIT) { @@ -1574,7 +1576,7 @@ split_fallthrough: pte = *ptep; if (!pte_present(pte)) goto no_page; - if ((flags & FOLL_NUMA) && pte_prot_none(vma, pte)) + if ((flags & FOLL_NUMA) && pte_numa(vma, pte)) goto no_page; if ((flags & FOLL_WRITE) && !pte_write(pte)) goto unlock; @@ -3476,7 +3478,7 @@ static int do_nonlinear_fault(struct mm_struct *mm, struct vm_area_struct *vma, return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte); } -static int do_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +static int do_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *ptep, pmd_t *pmd, unsigned int flags, pte_t entry) { @@ -3601,8 +3603,8 @@ int handle_pte_fault(struct mm_struct *mm, pte, pmd, flags, entry); } - if (pte_prot_none(vma, entry)) - return do_prot_none(mm, vma, address, pte, pmd, flags, entry); + if (pte_numa(vma, entry)) + return do_numa_page(mm, vma, address, pte, pmd, flags, entry); ptl = pte_lockptr(mm, pmd); spin_lock(ptl); @@ -3672,8 +3674,8 @@ retry: barrier(); if (pmd_trans_huge(orig_pmd) && !pmd_trans_splitting(orig_pmd)) { - if (pmd_prot_none(vma, orig_pmd)) { - do_huge_pmd_prot_none(mm, vma, address, pmd, + if (pmd_numa(vma, orig_pmd)) { + do_huge_pmd_numa_page(mm, vma, address, pmd, flags, orig_pmd); } ^ permalink raw reply related [flat|nested] 20+ messages in thread
* [tip:numa/core] numa, mm: Rename the PROT_NONE fault handling functions to *_numa() 2012-10-18 21:20 ` [PATCH 2/2] rename NUMA fault handling functions Rik van Riel 2012-10-19 11:41 ` Peter Zijlstra 2012-10-21 15:20 ` [tip:numa/core] numa, mm: Rename the PROT_NONE fault handling functions to *_numa() tip-bot for Rik van Riel @ 2012-10-23 11:00 ` tip-bot for Rik van Riel 2 siblings, 0 replies; 20+ messages in thread From: tip-bot for Rik van Riel @ 2012-10-23 11:00 UTC (permalink / raw) To: linux-tip-commits; +Cc: linux-kernel, riel, hpa, mingo, tglx Commit-ID: b3c01da073d82c8aaf3aa12f6214b64d2d1d83f8 Gitweb: http://git.kernel.org/tip/b3c01da073d82c8aaf3aa12f6214b64d2d1d83f8 Author: Rik van Riel <riel@redhat.com> AuthorDate: Thu, 18 Oct 2012 17:20:21 -0400 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Tue, 23 Oct 2012 11:53:51 +0200 numa, mm: Rename the PROT_NONE fault handling functions to *_numa() Having the function name indicate what the function is used for makes the code a little easier to read. Furthermore, the fault handling code largely consists of do_...._page functions. Rename the NUMA working set sampling fault handling functions to _numa() names, to indicate what they are used for. This separates the naming from the regular PROT_NONE namings. Signed-off-by: Rik van Riel <riel@redhat.com> Cc: aarcange@redhat.com Cc: a.p.zijlstra@chello.nl Link: http://lkml.kernel.org/r/20121018172021.0b1f6e3d@cuia.bos.redhat.com [ Converted two more usage sites ] Signed-off-by: Ingo Molnar <mingo@kernel.org> --- include/linux/huge_mm.h | 8 ++++---- mm/huge_memory.c | 4 ++-- mm/memory.c | 22 ++++++++++++---------- 3 files changed, 18 insertions(+), 16 deletions(-) diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h index bcbe467..4f0f948 100644 --- a/include/linux/huge_mm.h +++ b/include/linux/huge_mm.h @@ -160,9 +160,9 @@ static inline struct page *compound_trans_head(struct page *page) return page; } -extern bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd); +extern bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd); -extern void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +extern void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t orig_pmd); @@ -203,12 +203,12 @@ static inline int pmd_trans_huge_lock(pmd_t *pmd, return 0; } -static inline bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) +static inline bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd) { return false; } -static inline void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +static inline void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t orig_pmd) { diff --git a/mm/huge_memory.c b/mm/huge_memory.c index e62d3c5..bcba184 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -726,7 +726,7 @@ out: return handle_pte_fault(mm, vma, address, pte, pmd, flags); } -bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) +bool pmd_numa(struct vm_area_struct *vma, pmd_t pmd) { /* * See pte_prot_none(). @@ -737,7 +737,7 @@ bool pmd_prot_none(struct vm_area_struct *vma, pmd_t pmd) return pmd_same(pmd, pmd_modify(pmd, vma_prot_none(vma))); } -void do_huge_pmd_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +void do_huge_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pmd_t *pmd, unsigned int flags, pmd_t entry) { diff --git a/mm/memory.c b/mm/memory.c index b609354..7ff1905 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1471,11 +1471,13 @@ int zap_vma_ptes(struct vm_area_struct *vma, unsigned long address, } EXPORT_SYMBOL_GPL(zap_vma_ptes); -static bool pte_prot_none(struct vm_area_struct *vma, pte_t pte) +static bool pte_numa(struct vm_area_struct *vma, pte_t pte) { /* - * If we have the normal vma->vm_page_prot protections we're not a - * 'special' PROT_NONE page. + * For NUMA page faults, we use PROT_NONE ptes in VMAs with + * "normal" vma->vm_page_prot protections. Genuine PROT_NONE + * VMAs should never get here, because the fault handling code + * will notice that the VMA has no read or write permissions. * * This means we cannot get 'special' PROT_NONE faults from genuine * PROT_NONE maps, nor from PROT_WRITE file maps that do dirty @@ -1543,7 +1545,7 @@ struct page *follow_page(struct vm_area_struct *vma, unsigned long address, page = follow_huge_pmd(mm, address, pmd, flags & FOLL_WRITE); goto out; } - if ((flags & FOLL_NUMA) && pmd_prot_none(vma, *pmd)) + if ((flags & FOLL_NUMA) && pmd_numa(vma, *pmd)) goto no_page_table; if (pmd_trans_huge(*pmd)) { if (flags & FOLL_SPLIT) { @@ -1574,7 +1576,7 @@ split_fallthrough: pte = *ptep; if (!pte_present(pte)) goto no_page; - if ((flags & FOLL_NUMA) && pte_prot_none(vma, pte)) + if ((flags & FOLL_NUMA) && pte_numa(vma, pte)) goto no_page; if ((flags & FOLL_WRITE) && !pte_write(pte)) goto unlock; @@ -3476,7 +3478,7 @@ static int do_nonlinear_fault(struct mm_struct *mm, struct vm_area_struct *vma, return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte); } -static int do_prot_none(struct mm_struct *mm, struct vm_area_struct *vma, +static int do_numa_page(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long address, pte_t *ptep, pmd_t *pmd, unsigned int flags, pte_t entry) { @@ -3573,8 +3575,8 @@ int handle_pte_fault(struct mm_struct *mm, pte, pmd, flags, entry); } - if (pte_prot_none(vma, entry)) - return do_prot_none(mm, vma, address, pte, pmd, flags, entry); + if (pte_numa(vma, entry)) + return do_numa_page(mm, vma, address, pte, pmd, flags, entry); ptl = pte_lockptr(mm, pmd); spin_lock(ptl); @@ -3644,8 +3646,8 @@ retry: barrier(); if (pmd_trans_huge(orig_pmd) && !pmd_trans_splitting(orig_pmd)) { - if (pmd_prot_none(vma, orig_pmd)) { - do_huge_pmd_prot_none(mm, vma, address, pmd, + if (pmd_numa(vma, orig_pmd)) { + do_huge_pmd_numa_page(mm, vma, address, pmd, flags, orig_pmd); } ^ permalink raw reply related [flat|nested] 20+ messages in thread
* [PATCH 3/2] sched, numa, mm: Implement constant rate working set sampling 2012-10-18 21:18 [PATCH 0/2] minor NUMA cleanups & documentation Rik van Riel 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel 2012-10-18 21:20 ` [PATCH 2/2] rename NUMA fault handling functions Rik van Riel @ 2012-10-21 12:43 ` Ingo Molnar 2 siblings, 0 replies; 20+ messages in thread From: Ingo Molnar @ 2012-10-21 12:43 UTC (permalink / raw) To: Rik van Riel; +Cc: mingo, linux-kernel, aarcange, a.p.zijlstra * Rik van Riel <riel@redhat.com> wrote: > Hi Ingo, > > Here are some minor NUMA cleanups to start with. > > I have some ideas for larger improvements and ideas to port > over from autonuma, but I got caught up in some of the code > and am not sure about those changes yet. To help out I picked up a couple of obvious ones: cee8868763f8 x86, mm: Prevent gcc to re-read the pagetables a860d4c7a1f4 mm: Check if PTE is already allocated during page fault e9fe72334fb0 numa, mm: Fix NUMA hinting page faults from gup/gup_fast I kept Andrea as the author, the patches needed only minimal adaptation. Plus I finally completed testing and applying Peter's constant-rate WSS patch: 3d049f8a5398 sched, numa, mm: Implement constant, per task Working Set Sampling (WSS) rate This is in part similar to AutoNUMA's hinting page fault rate limiting feature (pages_to_scan et al), and in part an improvement/extension of it. See the patch below for details. Let me know if you have any questions! Thanks, Ingo ----------------> >From 3d049f8a5398d0050ab9978b3ac67402f337390f Mon Sep 17 00:00:00 2001 From: Peter Zijlstra <a.p.zijlstra@chello.nl> Date: Sun, 14 Oct 2012 16:59:13 +0200 Subject: [PATCH] sched, numa, mm: Implement constant, per task Working Set Sampling (WSS) rate Previously, to probe the working set of a task, we'd use a very simple and crude method: mark all of its address space PROT_NONE. That method has various (obvious) disadvantages: - it samples the working set at dissimilar rates, giving some tasks a sampling quality advantage over others. - creates performance problems for tasks with very large working sets - over-samples processes with large address spaces but which only very rarely execute Improve that method by keeping a rotating offset into the address space that marks the current position of the scan, and advance it by a constant rate (in a CPU cycles execution proportional manner). If the offset reaches the last mapped address of the mm then it then it starts over at the first address. The per-task nature of the working set sampling functionality in this tree allows such constant rate, per task, execution-weight proportional sampling of the working set, with an adaptive sampling interval/frequency that goes from once per 100 msecs up to just once per 1.6 seconds. The current sampling volume is 256 MB per interval. As tasks mature and converge their working set, so does the sampling rate slow down to just a trickle, 256 MB per 1.6 seconds of CPU time executed. This, beyond being adaptive, also rate-limits rarely executing systems and does not over-sample on overloaded systems. [ In AutoNUMA speak, this patch deals with the effective sampling rate of the 'hinting page fault'. AutoNUMA's scanning is currently rate-limited, but it is also fundamentally single-threaded, executing in the knuma_scand kernel thread, so the limit in AutoNUMA is global and does not scale up with the number of CPUs, nor does it scan tasks in an execution proportional manner. So the idea of rate-limiting the scanning was first implemented in the AutoNUMA tree via a global rate limit. This patch goes beyond that by implementing an execution rate proportional working set sampling rate that is not implemented via a single global scanning daemon. ] Based-on-idea-by: Andrea Arcangeli <aarcange@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Rik van Riel <riel@redhat.com> Link: http://lkml.kernel.org/n/tip-wt5b48o2226ec63784i58s3j@git.kernel.org [ Wrote changelog and fixed bug. ] Signed-off-by: Ingo Molnar <mingo@kernel.org> --- include/linux/mempolicy.h | 2 -- include/linux/mm.h | 6 ++++++ include/linux/mm_types.h | 1 + include/linux/sched.h | 1 + kernel/sched/fair.c | 44 ++++++++++++++++++++++++++++++++++++++++---- kernel/sysctl.c | 7 +++++++ mm/mempolicy.c | 24 ------------------------ 7 files changed, 55 insertions(+), 30 deletions(-) diff --git a/include/linux/mempolicy.h b/include/linux/mempolicy.h index a5bf9d6..d6b1ea1 100644 --- a/include/linux/mempolicy.h +++ b/include/linux/mempolicy.h @@ -199,8 +199,6 @@ static inline int vma_migratable(struct vm_area_struct *vma) extern int mpol_misplaced(struct page *, struct vm_area_struct *, unsigned long); -extern void lazy_migrate_process(struct mm_struct *mm); - #else /* CONFIG_NUMA */ struct mempolicy {}; diff --git a/include/linux/mm.h b/include/linux/mm.h index 423464b..64ccf29 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1581,6 +1581,12 @@ static inline pgprot_t vma_prot_none(struct vm_area_struct *vma) return pgprot_modify(vma->vm_page_prot, vm_get_page_prot(vmflags)); } +static inline void +change_prot_none(struct vm_area_struct *vma, unsigned long start, unsigned long end) +{ + change_protection(vma, start, end, vma_prot_none(vma), 0); +} + struct vm_area_struct *find_extend_vma(struct mm_struct *, unsigned long addr); int remap_pfn_range(struct vm_area_struct *, unsigned long addr, unsigned long pfn, unsigned long size, pgprot_t); diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index bef4c5e..01c1d04 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -405,6 +405,7 @@ struct mm_struct { #endif #ifdef CONFIG_SCHED_NUMA unsigned long numa_next_scan; + unsigned long numa_scan_offset; int numa_scan_seq; #endif struct uprobes_state uprobes_state; diff --git a/include/linux/sched.h b/include/linux/sched.h index 9e726f0..63c011e 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -2022,6 +2022,7 @@ extern enum sched_tunable_scaling sysctl_sched_tunable_scaling; extern unsigned int sysctl_sched_numa_task_period_min; extern unsigned int sysctl_sched_numa_task_period_max; +extern unsigned int sysctl_sched_numa_scan_size; extern unsigned int sysctl_sched_numa_settle_count; #ifdef CONFIG_SCHED_DEBUG diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index a66a1b6..9f7406e 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -827,8 +827,9 @@ static void account_numa_dequeue(struct rq *rq, struct task_struct *p) /* * numa task sample period in ms: 5s */ -unsigned int sysctl_sched_numa_task_period_min = 5000; -unsigned int sysctl_sched_numa_task_period_max = 5000*16; +unsigned int sysctl_sched_numa_task_period_min = 100; +unsigned int sysctl_sched_numa_task_period_max = 100*16; +unsigned int sysctl_sched_numa_scan_size = 256; /* MB */ /* * Wait for the 2-sample stuff to settle before migrating again @@ -902,6 +903,9 @@ void task_numa_work(struct callback_head *work) unsigned long migrate, next_scan, now = jiffies; struct task_struct *p = current; struct mm_struct *mm = p->mm; + struct vm_area_struct *vma; + unsigned long offset, end; + long length; WARN_ON_ONCE(p != container_of(work, struct task_struct, numa_work)); @@ -928,8 +932,40 @@ void task_numa_work(struct callback_head *work) if (cmpxchg(&mm->numa_next_scan, migrate, next_scan) != migrate) return; - ACCESS_ONCE(mm->numa_scan_seq)++; - lazy_migrate_process(mm); + + offset = mm->numa_scan_offset; + length = sysctl_sched_numa_scan_size; + length <<= 20; + + down_read(&mm->mmap_sem); + vma = find_vma(mm, offset); +again: + if (!vma) { + ACCESS_ONCE(mm->numa_scan_seq)++; + offset = 0; + vma = mm->mmap; + } + while (vma && !vma_migratable(vma)) { + vma = vma->vm_next; + if (!vma) + goto again; + } + + offset = max(offset, vma->vm_start); + end = min(ALIGN(offset + length, HPAGE_SIZE), vma->vm_end); + length -= end - offset; + + change_prot_none(vma, offset, end); + + offset = end; + + if (length > 0) { + vma = vma->vm_next; + goto again; + } + mm->numa_scan_offset = offset; + up_read(&mm->mmap_sem); + } /* diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 14a1949..0f0cb60 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -365,6 +365,13 @@ static struct ctl_table kern_table[] = { .proc_handler = proc_dointvec, }, { + .procname = "sched_numa_scan_size_mb", + .data = &sysctl_sched_numa_scan_size, + .maxlen = sizeof(unsigned int), + .mode = 0644, + .proc_handler = proc_dointvec, + }, + { .procname = "sched_numa_settle_count", .data = &sysctl_sched_numa_settle_count, .maxlen = sizeof(unsigned int), diff --git a/mm/mempolicy.c b/mm/mempolicy.c index f0e3b28..d998810 100644 --- a/mm/mempolicy.c +++ b/mm/mempolicy.c @@ -581,12 +581,6 @@ static inline int check_pgd_range(struct vm_area_struct *vma, return 0; } -static void -change_prot_none(struct vm_area_struct *vma, unsigned long start, unsigned long end) -{ - change_protection(vma, start, end, vma_prot_none(vma), 0); -} - /* * Check if all pages in a range are on a set of nodes. * If pagelist != NULL then isolate pages from the LRU and @@ -1259,24 +1253,6 @@ static long do_mbind(unsigned long start, unsigned long len, return err; } -static void lazy_migrate_vma(struct vm_area_struct *vma) -{ - if (!vma_migratable(vma)) - return; - - change_prot_none(vma, vma->vm_start, vma->vm_end); -} - -void lazy_migrate_process(struct mm_struct *mm) -{ - struct vm_area_struct *vma; - - down_read(&mm->mmap_sem); - for (vma = mm->mmap; vma; vma = vma->vm_next) - lazy_migrate_vma(vma); - up_read(&mm->mmap_sem); -} - /* * User space interface with variable sized bitmaps for nodelists. */ ^ permalink raw reply related [flat|nested] 20+ messages in thread
end of thread, other threads:[~2012-11-19 19:47 UTC | newest] Thread overview: 20+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2012-10-18 21:18 [PATCH 0/2] minor NUMA cleanups & documentation Rik van Riel 2012-10-18 21:19 ` [PATCH 1/2] add credits for NUMA placement Rik van Riel 2012-10-19 11:41 ` Peter Zijlstra 2012-10-19 12:02 ` [tip:numa/core] numa: Add " tip-bot for Rik van Riel 2012-10-28 17:11 ` [tip:numa/core] sched, numa, mm: " tip-bot for Rik van Riel 2012-11-13 15:31 ` tip-bot for Rik van Riel 2012-11-13 17:23 ` tip-bot for Rik van Riel 2012-11-19 19:46 ` tip-bot for Rik van Riel 2012-10-18 21:20 ` [PATCH 2/2] rename NUMA fault handling functions Rik van Riel 2012-10-19 11:41 ` Peter Zijlstra 2012-10-19 14:07 ` Rik van Riel 2012-10-19 20:54 ` Ingo Molnar 2012-10-21 12:50 ` [PATCH 4/2] numa, mm: Rename the PROT_NONE " Ingo Molnar 2012-10-21 13:22 ` Rik van Riel 2012-10-21 13:29 ` Ingo Molnar 2012-10-21 13:43 ` Ingo Molnar 2012-10-20 10:15 ` [PATCH 2/2] rename NUMA " Michel Lespinasse 2012-10-21 15:20 ` [tip:numa/core] numa, mm: Rename the PROT_NONE fault handling functions to *_numa() tip-bot for Rik van Riel 2012-10-23 11:00 ` tip-bot for Rik van Riel 2012-10-21 12:43 ` [PATCH 3/2] sched, numa, mm: Implement constant rate working set sampling Ingo Molnar
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).