[PATCH 11/19] sched/numa: Restrict migrating in parallel to the same node.

Sys: 319.46 418.57 363.31 37.62 -29.47% numa05.sh User:33727.7734732.6834127.41 447.11 -1.353% The commit does cause some performance regression but is needed from a fairness/correctness perspective. Signed-off-by: Srikar Dronamraju --- include

[PATCH 05/19] sched/numa: Use task faults only if numa_group is not yet setup

36654.5335074.51 1187.71 3.368% Ideally this change shouldn't have affected performance. Signed-off-by: Srikar Dronamraju --- kernel/sched/fair.c | 20 ++-- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c

[PATCH 03/19] sched/numa: Simplify load_too_imbalanced

25.5436896.3135637.84 1222.64 -2.12% Signed-off-by: Srikar Dronamraju --- kernel/sched/fair.c | 20 ++-- 1 file changed, 2 insertions(+), 18 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 57d1ee8..ea32a66 100644 --- a/kernel/sched/fair.c +++ b/k

[PATCH 04/19] sched/numa: Set preferred_node based on best_cpu

4732.1238016.8036255.85 1070.51 -1.704% While there is a performance hit, this is a correctness issue that is very much needed in bigger systems. Signed-off-by: Srikar Dronamraju --- kernel/sched/fair.c | 13 ++--- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/ke

[PATCH 02/19] sched/numa: Evaluate move once per node

% numa05.sh User:33255.8636890.8234879.31 1641.98 12.11% Signed-off-by: Srikar Dronamraju --- kernel/sched/fair.c | 128 +++- 1 file changed, 57 insertions(+), 71 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched

Re: [PATCH 04/19] sched/numa: Set preferred_node based on best_cpu

* Peter Zijlstra [2018-06-04 14:18:00]: > On Mon, Jun 04, 2018 at 03:30:13PM +0530, Srikar Dronamraju wrote: > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c > > index ea32a66..94091e6 100644 > > --- a/kernel/sched/fair.c > > +++ b/kernel/sched/fair.c >

Re: [PATCH 04/19] sched/numa: Set preferred_node based on best_cpu

* Peter Zijlstra [2018-06-04 14:23:36]: > OK, the above matches the description, but I'm puzzled by the remainder: > > > > > - if (ng->active_nodes > 1 && numa_is_active_node(env.dst_nid, > > ng)) > > - sched_setnuma(p, env.dst_nid); > > + if (nid != p->numa

Re: [PATCH 05/19] sched/numa: Use task faults only if numa_group is not yet setup

> > Testcase Time: Min Max Avg StdDev > > %Change > > numa01.sh Real: 478.45 565.90 515.11 30.87 > > 16.29% > > numa01.sh Sys: 207.79 271.04 232.94 21.33 > > -15.8% > > numa01.sh

Re: [PATCH 04/19] sched/numa: Set preferred_node based on best_cpu

* Peter Zijlstra [2018-06-04 15:39:53]: > > > > > > That seems to entirely loose the active_node thing, or are you saying > > > best_cpu already includes that? (Changelog could use a little help there > > > I suppose) > > > > I think checking for active_nodes before calling sched_setnuma was a >

Re: [PATCH 02/19] sched/numa: Evaluate move once per node

* Rik van Riel [2018-06-04 10:51:27]: > On Mon, 2018-06-04 at 15:30 +0530, Srikar Dronamraju wrote: > > > Just bike shedding, but it may be easier to read > if the "we found our destination" check were written > more explicitly: > > > if (!cur) {

Re: [PATCH 04/19] sched/numa: Set preferred_node based on best_cpu

* Rik van Riel [2018-06-04 10:37:30]: > On Mon, 2018-06-04 at 05:59 -0700, Srikar Dronamraju wrote: > > * Peter Zijlstra [2018-06-04 14:23:36]: > > > > > > - if (ng->active_nodes > 1 && > > > > numa_is_active_node(env.dst_nid,

Re: [PATCH 16/19] sched/numa: Detect if node actively handling migration

* Rik van Riel [2018-06-04 16:05:55]: > On Mon, 2018-06-04 at 15:30 +0530, Srikar Dronamraju wrote: > > > @@ -1554,6 +1562,9 @@ static void task_numa_compare(struct > > task_numa_env *env, > > if (READ_ONCE(dst_rq->numa_migrate_on)) > >

Re: [PATCH 13/19] mm/migrate: Use xchg instead of spinlock

* Peter Zijlstra [2018-06-04 21:28:21]: > > if (time_after(jiffies, pgdat->numabalancing_migrate_next_window)) { > > - spin_lock(&pgdat->numabalancing_migrate_lock); > > - pgdat->numabalancing_migrate_nr_pages = 0; > > - pgdat->numabalancing_migrate_next_window =

Re: [PATCH 16/19] sched/numa: Detect if node actively handling migration

> > > > I thought about this. Lets say we evaluated that destination node can > > allow movement. While we iterate through the list of cpus trying to > > find > > the best cpu node, we find a idle cpu towards the end of the list. > > However if another task as already raced with us to move a task

Re: [PATCH 11/19] sched/numa: Restrict migrating in parallel to the same node.

> > The commit does cause some performance regression but is needed from > > a fairness/correctness perspective. > > > > While it may cause some performance regressions, it may be due to either > a) some workloads benefit from overloading a node if the tasks idle > frequently or b) the regression

Re: [PATCH 18/19] sched/numa: Reset scan rate whenever task moves across nodes

* Mel Gorman [2018-06-05 10:58:43]: > On Mon, Jun 04, 2018 at 03:30:27PM +0530, Srikar Dronamraju wrote: > > Currently task scan rate is reset when numa balancer migrates the task > > to a different node. If numa balancer initiates a swap, reset is only > > applicable to th

Re: [PATCH 16/19] sched/numa: Detect if node actively handling migration

> > > > All tasks will not be stuck at task/cpu A. > > > > "[PATCH 10/19] sched/numa: Stop multiple tasks from moving to the > > cpu..." the first task to set cpu A as swap target will ensure > > subsequent tasks wont be allowed to set cpu A as target for swap till > > it > > finds a better task/

Re: [PATCH] sched/topology: Switch to sched_debug() for conditional sched domain printk

hed_debug()) { Same as above. > pr_info("root domain span: %*pbl (max cpu_capacity = %lu)\n", > cpumask_pr_args(cpu_map), rq->rd->max_cpu_capacity); > } > -- > 1.9.1 > -- Thanks and Regards Srikar Dronamraju

Re: [PATCH] sched/topology: Switch to sched_debug() for conditional sched domain printk

* Yue Hu [2021-02-03 18:10:19]: > On Wed, 3 Feb 2021 15:22:56 +0530 > Srikar Dronamraju wrote: > > > * Yue Hu [2021-02-03 12:20:10]: > > > > > > sched_debug() would only be present in CONFIG_SCHED_DEBUG. Right? > > In which case there would

Re: [PATCH] perf bench numa: Fix the condition checks for max number of numa nodes

ailed. > Aborted (core dumped) > <<>> > Looks good to me. Reviewed-by: Srikar Dronamraju -- Thanks and Regards Srikar Dronamraju

[PATCH] sched/fair: Prefer idle CPU to cache affinity

r Eggemann Cc: Mel Gorman Cc: Vincent Guittot Co-developed-by: Gautham R Shenoy Signed-off-by: Gautham R Shenoy Co-developed-by: Parth Shah Signed-off-by: Parth Shah Signed-off-by: Srikar Dronamraju --- kernel/sched/fair.c | 41 +++-- kernel/sched/featu

Re: [PATCH v4 00/10] Coregroup support on Powerpc

* Srikar Dronamraju [2020-07-27 11:02:20]: > Changelog v3 ->v4: > v3: > https://lore.kernel.org/lkml/20200723085116.4731-1-sri...@linux.vnet.ibm.com/t/#u > Here is a summary of some of the testing done with coregroup v4 patchsets. It includes ebizzy, schbench, perf bench

Re: [PATCH v4 07/10] Powerpc/numa: Detect support for coregroup

* Michael Ellerman [2020-07-31 17:49:55]: > Srikar Dronamraju writes: > > Add support for grouping cores based on the device-tree classification. > > - The last domain in the associativity domains always refers to the > > core. > > - If primary reference domain ha

Re: [PATCH v4 06/10] powerpc/smp: Generalize 2nd sched domain

* Michael Ellerman [2020-07-31 17:45:37]: > Srikar Dronamraju writes: > > Currently "CACHE" domain happens to be the 2nd sched domain as per > > powerpc_topology. This domain will collapse if cpumask of l2-cache is > > same as SMT domain. However we could gene

Re: [PATCH v4 08/10] powerpc/smp: Allocate cpumask only after searching thread group

* Michael Ellerman [2020-07-31 17:52:15]: > Srikar Dronamraju writes: > > If allocated earlier and the search fails, then cpumask need to be > > freed. However cpu_l1_cache_map can be allocated after we search thread > > group. > > It's not freed anywhere

Re: [PATCH v4 10/10] powerpc/smp: Implement cpu_to_coregroup_id

* Michael Ellerman [2020-07-31 18:02:21]: > Srikar Dronamraju writes: > > Lookup the coregroup id from the associativity array. > Thanks Michael for all your comments and inputs. > It's slightly strange that this is called in patch 9, but only properly > imple

Re: [PATCH v5 3/3] mm/page_alloc: Keep memoryless cpuless node 0 offline

Hi Andrew, Michal, David * Andrew Morton [2020-08-06 21:32:11]: > On Fri, 3 Jul 2020 18:28:23 +0530 Srikar Dronamraju > wrote: > > > > The memory hotplug changes that somehow because you can hotremove numa > > > nodes and therefore make the nodemask sparse but that

Re: [PATCH v4 09/10] Powerpc/smp: Create coregroup domain

eflect your LLC situation via this > flag to make cpus_share_cache() work properly. I detect if the LLC is shared at BIGCORE, and if they are shared at BIGCORE, then I dynamically rename the DOMAIN as CACHE and enable SD_SHARE_PKG_RESOURCES in that domain. > > [1]: https://linuxplumbersconf.org/event/4/contributions/484/ Thanks for the pointer. -- Thanks and Regards Srikar Dronamraju

Re: [PATCH 1/2] sched/topology: Allow archs to override cpu_smt_mask

Zijlstra (Intel) > > An updated Changelog that recaps some of this discussion might also be > nice. Okay, will surely do the needful. -- Thanks and Regards Srikar Dronamraju

Re: [PATCH 1/2] sched/topology: Allow archs to override cpu_smt_mask

of reboot they would only have the older P8 topology. After reboot the kernel topology would change, but the userspace is made to believe that they are running on SMT8 core by way of keeping the sibling_cpumask at SMT8 core level. -- Thanks and Regards Srikar Dronamraju

[PATCH v2 2/2] powerpc/topology: Override cpu_smt_mask

: LKML Cc: Michael Ellerman Cc: Michael Neuling Cc: Gautham R Shenoy Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Dietmar Eggemann Cc: Mel Gorman Cc: Vincent Guittot Cc: Vaidyanathan Srinivasan Signed-off-by: Srikar Dronamraju --- Changelog v1->v2: Modified com

[PATCH v2 1/2] sched/topology: Allow archs to override cpu_smt_mask

uling Cc: Gautham R Shenoy Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Dietmar Eggemann Cc: Mel Gorman Cc: Vincent Guittot Cc: Vaidyanathan Srinivasan Acked-by; Peter Zijlstra (Intel) Signed-off-by: Srikar Dronamraju --- Changelog v1->v2: Update the commit msg

Re: [PATCH v2] sched/fair: ignore cache hotness for SMT migration

SD_SHARE_CPUCAPACITY? > /* >* Buddy candidates are cache hot: > */ > -- > 2.28.0.163.g6104cc2f0b6-goog > -- Thanks and Regards Srikar Dronamraju

[PATCH v5 04/10] powerpc/smp: Move topology fixups into a new function

henoy Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Jordan Niethe Cc: Vaidyanathan Srinivasan Reviewed-by: Gautham R. Shenoy Signed-off-by: Srikar Dronamraju --- Changelog v2 -> v3: Rewrote changelog (Gautham) Renamed to powerpc/smp: Move topology fixups

[PATCH v5 06/10] powerpc/smp: Optimize start_secondary

Ellerman Cc: Nicholas Piggin Cc: Anton Blanchard Cc: Oliver O'Halloran Cc: Nathan Lynch Cc: Michael Neuling Cc: Gautham R Shenoy Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Jordan Niethe Cc: Vaidyanathan Srinivasan Signed-off-by: Srikar Dronamraju --- Changelog v4

[PATCH v5 10/10] powerpc/smp: Implement cpu_to_coregroup_id

ned-off-by: Srikar Dronamraju --- Changelog v1 -> v2: Move coregroup_enabled before getting associativity (Gautham) arch/powerpc/mm/numa.c | 20 1 file changed, 20 insertions(+) diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c index 0d57779e7942..8b3b3e

[PATCH v5 05/10] powerpc/smp: Dont assume l2-cache to be superset of sibling

Piggin Cc: Anton Blanchard Cc: Oliver O'Halloran Cc: Nathan Lynch Cc: Michael Neuling Cc: Gautham R Shenoy Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Jordan Niethe Cc: Vaidyanathan Srinivasan Reviewed-by: Gautham R. Shenoy Signed-off-by: Srikar Dronamraju --- Changel

[PATCH v5 00/10] Coregroup support on Powerpc

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Cc: linuxppc-dev Cc: LKML Cc: Michael Ellerman Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Nick Piggin Cc: Oliver OHalloran Cc: Nathan Lynch Cc: Michael Neuling Cc: Anton Blanchard Cc: Gautham R Shenoy Cc: Vaidyan

[PATCH v5 09/10] powerpc/smp: Create coregroup domain

Srinivasan Reviewed-by: Gautham R. Shenoy Signed-off-by: Srikar Dronamraju --- Changelog v4 ->v5: Updated commit msg to specify actual implementation of cpu_to_coregroup_id is in a subsequent patch (Michael Ellerman) Changelog v3 ->v4: if coregroup_support doesn&#x

[PATCH v5 08/10] powerpc/smp: Allocate cpumask only after searching thread group

vasan Reviewed-by: Gautham R. Shenoy Signed-off-by: Srikar Dronamraju --- Changelog v4 ->v5: Updated commit msg on why cpumask need not be freed. (Michael Ellerman) arch/powerpc/kernel/smp.c | 7 +++ 1 file changed, 3 insertions(+), 4 d

[PATCH v5 07/10] powerpc/numa: Detect support for coregroup

h Cc: Michael Neuling Cc: Gautham R Shenoy Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Jordan Niethe Cc: Vaidyanathan Srinivasan Reviewed-by: Gautham R. Shenoy Signed-off-by: Srikar Dronamraju --- Changelog v4->v5: Updated commit msg with current abstract natur

[PATCH v5 03/10] powerpc/smp: Move powerpc_topology above

chael Neuling Cc: Gautham R Shenoy Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Jordan Niethe Cc: Vaidyanathan Srinivasan Reviewed-by: Gautham R. Shenoy Signed-off-by: Srikar Dronamraju --- arch/powerpc/kernel/smp.c | 104 +++--- 1 file ch

[PATCH v5 01/10] powerpc/smp: Fix a warning under !NEED_MULTIPLE_NODES

Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Jordan Niethe Cc: Vaidyanathan Srinivasan Reviewed-by: Gautham R. Shenoy Signed-off-by: Srikar Dronamraju --- Changelog v2 -> v3: Removed node caching part. Rewrote the Commit msg (Michael Ellerman) Renamed to powerpc/

[PATCH v5 02/10] powerpc/smp: Merge Power9 topology with Power topology

: Anton Blanchard Cc: Oliver O'Halloran Cc: Nathan Lynch Cc: Michael Neuling Cc: Gautham R Shenoy Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Jordan Niethe Cc: Vaidyanathan Srinivasan Reviewed-by: Gautham R. Shenoy Signed-off-by: Srikar Dronamraju --- Changelog v1

Re: [PATCH v5 3/3] mm/page_alloc: Keep memoryless cpuless node 0 offline

* Michal Hocko [2020-08-18 09:37:12]: > On Tue 18-08-20 09:32:52, David Hildenbrand wrote: > > On 12.08.20 08:01, Srikar Dronamraju wrote: > > > Hi Andrew, Michal, David > > > > > > * Andrew Morton [2020-08-06 21:32:11]: > > > > > >

Re: [PATCH 1/3] powerpc/smp: Parse ibm,thread-groups with multiple properties

int i_group_start = get_cpu_thread_group_start(i, tg); > > if (unlikely(i_group_start == -1)) { > WARN_ON_ONCE(1); > @@ -843,7 +881,7 @@ static int init_cpu_l1_cache_map(int cpu) > } > > if (i_group_start == cpu_group_start) > - cpumask_set_cpu(i, per_cpu(cpu_l1_cache_map, cpu)); > + cpumask_set_cpu(i, *mask); > } > > out: > @@ -924,7 +962,7 @@ static int init_big_cores(void) > int cpu; > > for_each_possible_cpu(cpu) { > - int err = init_cpu_l1_cache_map(cpu); > + int err = init_cpu_cache_map(cpu, THREAD_GROUP_SHARE_L1); > > if (err) > return err; > -- > 1.9.4 > -- Thanks and Regards Srikar Dronamraju

Re: [PATCH 2/3] powerpc/smp: Add support detecting thread-groups sharing L2 cache

for_each_cpu(i, *mask) { > + if (!cpu_online(i)) > + continue; > + set_cpus_related(i, cpu, cpu_l2_cache_mask); > + } > + > + return true; > + } > + Ah this can be simplified to: if (thread_group_shares_l2) { cpumask_set_cpu(cpu, cpu_l2_cache_mask(cpu)); for_each_cpu(i, per_cpu(thread_group_l2_cache_map, cpu)) { if (cpu_online(i)) set_cpus_related(i, cpu, cpu_l2_cache_mask); } } No? > l2_cache = cpu_to_l2cache(cpu); > if (!l2_cache || !*mask) { > /* Assume only core siblings share cache with this CPU */ -- Thanks and Regards Srikar Dronamraju

Re: [PATCH 3/3] powerpc/cacheinfo: Print correct cache-sibling map/list for L2 cache

he first place. For example:- If for a P9 core with CPUs 0-7, the cache->shared_cpu_map for L1 would have 0-7 but would display 0,2,4,6. The drawback of this is even if cpus 0,2,4,6 are released L1 cache will not be released. Is this as expected? -- Thanks and Regards Srikar Dronamraju

[PATCH v2] sched/fair: Prefer idle CPU to cache affinity

Gautham R Shenoy Signed-off-by: Gautham R Shenoy Co-developed-by: Parth Shah Signed-off-by: Parth Shah Signed-off-by: Srikar Dronamraju --- Changelog v1->v2: v1: http://lore.kernel.org/lkml/20210226164029.122432-1-sri...@linux.vnet.ibm.com/t/#u - Make WA_WAKER default (Suggested by Rik) -

Re: [RFC/PATCH] powerpc/smp: Add SD_SHARE_PKG_RESOURCES flag to MC sched-domain

isphere, and finally across hemispheres), do you have any suggestions on how we could handle the same in the core scheduler? -- Thanks and Regards Srikar Dronamraju

Re: [PATCH 1/3] powerpc/smp: Parse ibm,thread-groups with multiple properties

> + zalloc_cpumask_var_node(mask, GFP_KERNEL, cpu_to_node(cpu)); > > > > > > > This hunk (and the next hunk) should be moved to next patch. > > > > The next patch is only about introducing THREAD_GROUP_SHARE_L2. Hence > I put in any other code in this

Re: [PATCH 3/3] powerpc/cacheinfo: Print correct cache-sibling map/list for L2 cache

just that there is still something more left to be done. -- Thanks and Regards Srikar Dronamraju

Re: [PATCH 2/3] powerpc/smp: Add support detecting thread-groups sharing L2 cache

n't we want to enforce that the siblings sharing L1 be a subset of > the siblings sharing L2 ? Or do you recommend putting in a check for > that somewhere ? > I didnt think about the case where the device-tree could show L2 to be a subset of L1. How about initializing thread_group_l2_cache_map itself with cpu_l1_cache_map. It would be a simple one time operation and reduce the overhead here every CPU online. And it would help in your subsequent patch too. We dont want the cacheinfo for L1 showing CPUs not present in L2. -- Thanks and Regards Srikar Dronamraju

Re: powerpc64-linux-ld: arch/powerpc/mm/numa.o:undefined reference to `coregroup_enabled'

7; > > powerpc64-linux-ld: mm/khugepaged.o:(.toc+0x0): undefined reference to > > `node_reclaim_distance' > > Hm, OK. > CONFIG_NUMA=y > # CONFIG_SMP is not set > > Michael, Gautham, does anyone care about this config combination? > I can add #ifdef CONFIG_SMP where coregroup_enabled is being accessed but I do feel CONFIG_NUMA but !CONFIG_SMP may not be a valid combination. > > Thanks. -- Thanks and Regards Srikar Dronamraju

[PATCH v2 1/4] powerpc: Refactor is_kvm_guest declaration to new header

Cc: Phil Auld Acked-by: Waiman Long Signed-off-by: Srikar Dronamraju --- Changelog: v1->v2: v1: https://lore.kernel.org/linuxppc-dev/20201028123512.871051-1-sri...@linux.vnet.ibm.com/t/#u - Moved a hunk to fix a no previous prototype warning reported by: l...@intel.com https://lists.01.

[PATCH v2 2/4] powerpc: Rename is_kvm_guest to check_kvm_guest

: Juri Lelli Cc: Waiman Long Cc: Phil Auld Acked-by: Waiman Long Signed-off-by: Srikar Dronamraju --- arch/powerpc/include/asm/kvm_guest.h | 4 ++-- arch/powerpc/include/asm/kvm_para.h | 2 +- arch/powerpc/kernel/firmware.c | 2 +- arch/powerpc/platforms/pseries/smp.c | 2 +- 4 files

[PATCH v2 0/4] Powerpc: Better preemption for shared processor

las Piggin Cc: Nathan Lynch Cc: Gautham R Shenoy Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Juri Lelli Cc: Waiman Long Cc: Phil Auld Srikar Dronamraju (4): powerpc: Refactor is_kvm_guest declaration to new header powerpc: Rename is_kvm_guest to check_kvm_guest powerpc: Reintrod

[PATCH v2 3/4] powerpc: Reintroduce is_kvm_guest in a new avatar

: Nicholas Piggin Cc: Nathan Lynch Cc: Gautham R Shenoy Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Juri Lelli Cc: Waiman Long Cc: Phil Auld Acked-by: Waiman Long Signed-off-by: Srikar Dronamraju --- arch/powerpc/include/asm/kvm_guest.h | 10 ++ arch/powerpc/include/asm

[PATCH v2 4/4] powerpc/paravirt: Use is_kvm_guest in vcpu_is_preempted

: Michael Ellerman Cc: Nicholas Piggin Cc: Nathan Lynch Cc: Gautham R Shenoy Cc: Peter Zijlstra Cc: Valentin Schneider Cc: Juri Lelli Cc: Waiman Long Cc: Phil Auld Acked-by: Waiman Long Signed-off-by: Srikar Dronamraju --- arch/powerpc/include/asm/paravirt.h | 18 ++ 1 file

Re: [PATCH 2/3] x86/uprobes: Fix not using prefixes.nbytes for loop over prefixes.bytes

; Fixes: 2b1444983508 ("uprobes, mm, x86: Add the ability to install and remove > uprobes breakpoints") > Cc: sta...@vger.kernel.org > Reported-by: Kees Cook > Signed-off-by: Masami Hiramatsu Looks good to me. Reviewed-by: Srikar Dronamraju > --- > arch/x86/kernel/upro

Re: [PATCH] sched/fair: Prefer idle CPU to cache affinity

we need to be conservative esp if we want to make WA_WAKER on by default. I would still like to hear from other people if they think its ok to enable it by default. I wonder if enabling it by default can cause some load imbalances leading to more active load balance down the line. I haven't benchmarked with WA_WAKER enabled. Thanks Rik for your inputs. -- Thanks and Regards Srikar Dronamraju

Re: [PATCH] sched/fair: Prefer idle CPU to cache affinity

* Peter Zijlstra [2021-03-01 16:44:42]: > On Sat, Feb 27, 2021 at 02:56:07PM -0500, Rik van Riel wrote: > > On Fri, 2021-02-26 at 22:10 +0530, Srikar Dronamraju wrote: > > > > + if (sched_feat(WA_WAKER) && tnr_busy < tllc_size) > > > + return

Re: [PATCH] sched/fair: Prefer idle CPU to cache affinity

* Peter Zijlstra [2021-03-01 16:40:33]: > On Fri, Feb 26, 2021 at 10:10:29PM +0530, Srikar Dronamraju wrote: > > +static int prefer_idler_llc(int this_cpu, int prev_cpu, int sync) > > +{ > > + struct sched_domain_shared *tsds, *psds; > > + int pnr_busy, pllc_size

Re: [PATCH] sched/fair: Prefer idle CPU to cache affinity

* Peter Zijlstra [2021-03-01 18:18:28]: > On Mon, Mar 01, 2021 at 10:36:01PM +0530, Srikar Dronamraju wrote: > > * Peter Zijlstra [2021-03-01 16:44:42]: > > > > > On Sat, Feb 27, 2021 at 02:56:07PM -0500, Rik van Riel wrote: > > > > On Fri, 2021-02-26 at

Re: [PATCH] sched/fair: Prefer idle CPU to cache affinity

* Dietmar Eggemann [2021-03-02 10:53:06]: > On 26/02/2021 17:40, Srikar Dronamraju wrote: > > [...] > > > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c > > index 8a8bd7b13634..d49bfcdc4a19 100644 > > --- a/kernel/sched/fair.c > > +++ b/kernel/

Re: [PATCH] sched/fair: Prefer idle CPU to cache affinity

* Peter Zijlstra [2021-03-02 10:10:32]: > On Tue, Mar 02, 2021 at 01:09:46PM +0530, Srikar Dronamraju wrote: > > > Oh, could be, I didn't grep :/ We could have core code keep track of the > > > smt count I suppose. > > > > Could we use cpumask_

Re: [PATCH] sched/fair: Prefer idle CPU to cache affinity

* Vincent Guittot [2021-03-08 14:52:39]: > On Fri, 26 Feb 2021 at 17:41, Srikar Dronamraju > wrote: > > Thanks Vincent for your review comments. > > +static int prefer_idler_llc(int this_cpu, int prev_cpu, int sync) > > +{ > > + struct sched_domain_shared

Re: [PATCH v3 2/5] powerpc/smp: Rename cpu_l1_cache_map as thread_group_l1_cache_map

nstruction and Data flow. > > This patch renames the variable to "thread_group_l1_cache_map" to make > it consistent with a subsequent patch which will introduce > thread_group_l2_cache_map. > > This patch introduces no functional change. > Looks good to me. R

Re: [PATCH v3 3/5] powerpc/smp: Rename init_thread_group_l1_cache_map() to make it generic

property (L1 or > L2) and update a suitable mask. This is a preparatory patch for the > next patch where we will introduce discovery of thread-groups that > share L2-cache. > > No functional change. > Looks good to me. Reviewed-by: Srikar Dronamraju > Signed-off-by: Ga

Re: [PATCH v3 4/5] powerpc/smp: Add support detecting thread-groups sharing L2 cache

00006 0001 0003 > 0005 0007 > Looks good to me. Reviewed-by: Srikar Dronamraju -- Thanks and Regards Srikar Dronamraju

Re: [PATCH v3 5/5] powerpc/cacheinfo: Print correct cache-sibling map/list for L2 cache

0004 0006 0001 > 0003 0005 0007 0002 > 0002 0004 0002 > 0004 0006 0001 0003 > 0005 0007 > Looks good to me. Reviewed-by: Srikar Dronamraju -- Thanks and Regards Srikar Dronamraju

Re: [PATCH v3 1/5] powerpc/smp: Parse ibm,thread-groups with multiple properties

y is being shared by > which groups of threads. This array can encode information about > multiple properties being shared by different thread-groups within the > core. > Looks good to me. Reviewed-by: Srikar Dronamraju -- Thanks and Regards Srikar Dronamraju

Re: [PATCH] sched: put rq's sched_avg under CONFIG_FAIR_GROUP_SCHED

e updated or used for rq->avg. Should we look at splitting sched_avg so that rq->avg doesn't have unwanted fields? -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.

Re: [PATCH 02/11] sched: remove rq->cpu_load[load_idx] array

> - load_idx = 0; > - > do { > struct sg_lb_stats *sgs = &tmp_sgs; > int local_group; The single line change in the previous patch gets removed here so why not club them. -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this

Re: [GIT PULL] uprobes: preparations for arm port

nction declarations out of arch > > Oleg Nesterov (3): > uprobes: Kill module_init() and module_exit() > uprobes: Introduce arch_uprobe->ixol > uprobes: Export write_opcode() as uprobe_write_opcode() > Acked-by: Srikar Dronamraju for this series. -- Thanks

Re: [PATCH v6 00/14] uprobes: Add uprobes support for ARM

> hammer). We don't need to do this on ARM, and we don't do it. The > result is that, unless PERF_EVENT is set separately, uprobes tends > not to build. I was lucking-out in my testing due to other default > config items turning on PERF_EVENT. > -- Thanks and Regards Srikar

Tasks stuck in futex code (in 3.14-rc6)

wake up was the commit thats causing the threads to be stuck in futex. I reverted b0c29f79ecea0b6fbcefc999e70f2843ae8306db on top of v3.14-rc6 and confirmed that reverting the commit solved the problem. -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send

Re: Tasks stuck in futex code (in 3.14-rc6)

owever if I set the constraint to core (which means running more instances of java), the problem is not seen. I kind of guess, the lesser the number of java instances the easier it is to reproduce. -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send the line "

Re: Tasks stuck in futex code (in 3.14-rc6)

S 3fff825f6044 0 14682 14076 0x0080 Is there any other information that I provide that can help? -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org Mo

Re: Tasks stuck in futex code (in 3.14-rc6)

lds/linux.git/commit/?id=b0c29f79ecea0b6fbcefc999 are the same. Or am I missing something? -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://v

Re: Tasks stuck in futex code (in 3.14-rc6)

13/12/19/624 > https://lkml.org/lkml/2013/12/19/630 I reverted commits 99b60ce6 and b0c29f79. Then applied the patches in the above url. The last one had a reject but it was pretty straightforward to resolve it. After this, specjbb completes. So reverting and applying v3 3/4 and 4/4 patches

Re: [PATCH 1/7] uprobes: Kill UPROBE_SKIP_SSTEP and can_skip_sstep()

(). > > ppc: Looks like, it can emulate almost everything. Does it >actually needs to record the fact that emulate_step() >failed? Hopefully not. But if yes, it can add the ppc- > specific flag into arch_uprobe. > > TODO: rename arch_uprob

Re: [PATCH 2/7] uprobes/x86: Fold prepare_fixups() into arch_uprobe_analyze_insn()

ood_insns_32 should depend > on CONFIG_X86_32/EMULATION > > - the usage of mm->context.ia32_compat looks wrong if the task > is TIF_X32. > > Signed-off-by: Oleg Nesterov Acked-by: Srikar Dronamraju -- Thanks and Regards Srikar Dronamraju

Re: [PATCH 4/7] uprobes/x86: Kill the "ia32_compat" check in handle_riprel_insn(), remove "mm" arg

> checks insn->x86_64. > > Also, remove the no longer needed "struct mm_struct *mm" argument and > the unnecessary "return" at the end. > > Signed-off-by: Oleg Nesterov Acked-by: Srikar Dronamraju -- Thanks and Regards Srikar Dronamraju -- To unsubscribe

Re: [PATCH 6/7] uprobes/x86: move the UPROBE_FIX_{RIP,IP,CALL} code at the end of pre/post hooks

() to use utask instead of autask, to > make the code more symmetrical with arch_uprobe_post_xol(). > > Signed-off-by: Oleg Nesterov Acked-by: Srikar Dronamraju -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send the line "unsubscribe linux-kernel"

Re: [PATCH 03/17] tracing/kprobes: Factor out struct trace_probe

* Namhyung Kim [2013-11-27 15:19:49]: > From: Namhyung Kim > > There are functions that can be shared to both of kprobes and uprobes. > Separate common data structure to struct trace_probe and use it from > the shared functions. > > Acked-by: Masami Hiramatsu > Cc:

Re: [PATCH 01/17] tracing/uprobes: Fix documentation of uprobe registration syntax

* Namhyung Kim [2013-11-27 15:19:47]: > From: Namhyung Kim > > The uprobe syntax requires an offset after a file path not a symbol. > > Reviewed-by: Masami Hiramatsu > Acked-by: Oleg Nesterov > Cc: Srikar Dronamraju > Cc: zhangwei(Jovi) > Cc: Arnaldo Carval

Re: [PATCH 04/17] tracing/uprobes: Convert to struct trace_probe

* Namhyung Kim [2013-11-27 15:19:50]: > From: Namhyung Kim > > Convert struct trace_uprobe to make use of the common trace_probe > structure. > > Reviewed-by: Masami Hiramatsu > Cc: Srikar Dronamraju > Cc: Oleg Nesterov > Cc: zhangwei(Jovi) > Cc: Arnaldo Car

Re: [PATCH] uprobes: Add uprobe_task->dup_work/dup_addr

hing like "private_data_for_handlers" so that the tracing > handlers could use it to communicate with call_fetch() methods. One nit below + request for change in the above commit message. Otherwise Acked-by: Srikar Dronamraju > > Signed-off-by: Oleg Nesterov > --- >

Re: [PATCH] uprobes: Cleanup !CONFIG_UPROBES decls, unexport xol_area

ys >nothing. > > 3. Kill the dummy definition of uprobe_get_swbp_addr(), nobody >except handle_swbp() needs it. > > 4. Purely cosmetic, but move the decl of uprobe_get_swbp_addr() >up, close to other __weak helpers. > > Signed-off-by: Oleg Nesterov Acked-by

Re: [PATCH 2/2] uprobes/powerpc: Kill arch_uprobe->ainsn

* Oleg Nesterov [2013-11-09 18:54:09]: > powerpc has both arch_uprobe->insn and arch_uprobe->ainsn to > make the generic code happy. This is no longer needed after > the previous change, powerpc can just use "u32 insn". > > Signed-off-by: Oleg Nesterov Acke

Re: [PATCH 1/2] uprobes: Don't assume that arch_uprobe->insn/ixol is u8[MAX_UINSN_BYTES]

me effect. > > Signed-off-by: Oleg Nesterov Acked-by: Srikar Dronamraju -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.

Re: [tip:sched/core] sched/fair: Fix group power_orig computation

8b 52 10 48 8b 52 10 8b 4a 08 8b 52 04 49 01 cc 48 01 d3 8 3 RIP [] update_group_power+0xa3/0x130 RSP CR2: 0010 ---[ end trace cd8cb7fb261d7bea ]--- Kernel panic - not syncing: Fatal exception This can be fixed by a simple check below. -- Thanks and Regards Srikar Dron

[PATCH v2] sched: Check sched_domain before computing group power.

After Commit-id 863bffc80898 (sched/fair: Fix group power_orig computation), we might end up computing group power before the sched_domain for a cpu is updated. Update with cpu_power, if rq->sd is not yet updated. Signed-off-by: Srikar Dronamraju --- Changelog since v1: Fix divide by z

Re: [PATCH v2] sched: Check sched_domain before computing group power.

> > Hurm.. can you provide the actual topology of the machine that triggers > this? My brain hurts trying to thing through the weird cases of this > code. > Hope this helps. Please do let me know if you were looking for pdf output. Machine (251GB) NUMANode P#0 (63GB) Socket P#0

Re: [PATCH] uprobes: Add uprobe_task->dup_work/dup_addr

Okay, moving to arch_uprobe_task is fine. I probably got confused by "First of all it is not really needed," > > > Your change still retains it. > > > OK. How about dup_xol_work/dup_xol_vaddr ? > Yes fine with me. -- Thanks and Regards Srikar Dronamraju -- To uns

Re: [PATCH v2] uprobes: Add uprobe_task->dup_work/dup_addr

* Oleg Nesterov [2013-11-12 20:20:38]: > On 11/12, Srikar Dronamraju wrote: > > > > Okay, moving to arch_uprobe_task is fine. I probably got confused by > > "First of all it is not really needed," > > OK, this doesn't look good, I agree. >

Re: [PATCH v2] sched: Check sched_domain before computing group power.

* Peter Zijlstra [2013-11-12 18:55:54]: > On Tue, Nov 12, 2013 at 10:45:07PM +0530, Srikar Dronamraju wrote: > > > > > > Hurm.. can you provide the actual topology of the machine that triggers > > > this? My brain hurts trying to thing through the

Re: [PATCH 1/1] uprobes: Fix the memory out of bound overwrite in copy_insn()

ote: we do not care if offset + size > i_size, the users of > arch_uprobe->insn can't know how many bytes were actually copied > anyway. But perhaps this needs more changes. > > Signed-off-by: Oleg Nesterov Acked-by: Srikar Dronamraju -- Thanks and Regards

Re: [PATCH v2] sched: Check sched_domain before computing group power.

roup_power(). > Please do elaborate on how you observed this. > Does this clarify? -- Thanks and Regards Srikar Dronamraju -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/

<    1   2   3   4   5   6   7   8   9   >