diff for duplicates of <540F4592.9030408@sr71.net> diff --git a/a/1.txt b/N1/1.txt index 8f3600f..fada128 100644 --- a/a/1.txt +++ b/N1/1.txt @@ -11,28 +11,28 @@ On 09/09/2014 07:50 AM, Johannes Weiner wrote: I got the wrong lock. Here's how it looks after mainline, plus your free_pages_and_swap_cache() patch: Samples: 2M of event 'cycles', Event count (approx.): 51647128377 -+ 60.60% 1.33% page_fault2_processes [.] testcase a?? -+ 59.14% 0.41% [kernel] [k] page_fault a?? -+ 58.72% 0.01% [kernel] [k] do_page_fault a?? -+ 58.70% 0.08% [kernel] [k] __do_page_fault a?? -+ 58.50% 0.29% [kernel] [k] handle_mm_fault a?? -+ 40.14% 0.28% [kernel] [k] do_cow_fault a?? -- 34.56% 34.56% [kernel] [k] _raw_spin_lock a?? - - _raw_spin_lock a?? - - 78.11% __res_counter_charge a?? - res_counter_charge a?? - try_charge a?? - - mem_cgroup_try_charge a?? - + 99.99% do_cow_fault a?? - - 10.30% res_counter_uncharge_until a?? - res_counter_uncharge a?? - uncharge_batch a?? - uncharge_list a?? - mem_cgroup_uncharge_list a?? - release_pages a?? - + 4.75% free_pcppages_bulk a?? - + 3.65% do_cow_fault a?? - + 2.24% get_page_from_freelist a?? ++ 60.60% 1.33% page_fault2_processes [.] testcase ▒ ++ 59.14% 0.41% [kernel] [k] page_fault ◆ ++ 58.72% 0.01% [kernel] [k] do_page_fault ▒ ++ 58.70% 0.08% [kernel] [k] __do_page_fault ▒ ++ 58.50% 0.29% [kernel] [k] handle_mm_fault ▒ ++ 40.14% 0.28% [kernel] [k] do_cow_fault ▒ +- 34.56% 34.56% [kernel] [k] _raw_spin_lock ▒ + - _raw_spin_lock ▒ + - 78.11% __res_counter_charge ▒ + res_counter_charge ▒ + try_charge ▒ + - mem_cgroup_try_charge ▒ + + 99.99% do_cow_fault ▒ + - 10.30% res_counter_uncharge_until ▒ + res_counter_uncharge ▒ + uncharge_batch ▒ + uncharge_list ▒ + mem_cgroup_uncharge_list ▒ + release_pages ▒ + + 4.75% free_pcppages_bulk ▒ + + 3.65% do_cow_fault ▒ + + 2.24% get_page_from_freelist ▒ > You also said that this cost hasn't been there before, but I do see > that trace in both v3.16 and v3.17-rc3 with roughly the same impact @@ -42,25 +42,19 @@ Samples: 2M of event 'cycles', Event count (approx.): 51647128377 Here's the same workload on the same machine with only Johannes' revert applied: -- 35.92% 35.92% [kernel] [k] _raw_spin_lock a?? - - _raw_spin_lock a?? - - 49.09% get_page_from_freelist a?? - - __alloc_pages_nodemask a?? - + 99.90% alloc_pages_vma a?? - - 43.67% free_pcppages_bulk a?? - - 100.00% free_hot_cold_page a?? - + 99.93% free_hot_cold_page_list a?? - - 7.08% do_cow_fault a?? - handle_mm_fault a?? - __do_page_fault a?? - do_page_fault a?? - page_fault a?? - testcase a?? +- 35.92% 35.92% [kernel] [k] _raw_spin_lock ▒ + - _raw_spin_lock ▒ + - 49.09% get_page_from_freelist ▒ + - __alloc_pages_nodemask ▒ + + 99.90% alloc_pages_vma ▒ + - 43.67% free_pcppages_bulk ▒ + - 100.00% free_hot_cold_page ▒ + + 99.93% free_hot_cold_page_list ▒ + - 7.08% do_cow_fault ▒ + handle_mm_fault ▒ + __do_page_fault ▒ + do_page_fault ▒ + page_fault ▒ + testcase ▒ So I think it's probably part of the same regression. - --- -To unsubscribe, send a message with 'unsubscribe linux-mm' in -the body to majordomo@kvack.org. For more info on Linux MM, -see: http://www.linux-mm.org/ . -Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a> diff --git a/a/content_digest b/N1/content_digest index c36ea51..9039cb0 100644 --- a/a/content_digest +++ b/N1/content_digest @@ -34,28 +34,28 @@ "I got the wrong lock. Here's how it looks after mainline, plus your free_pages_and_swap_cache() patch:\n" "\n" "Samples: 2M of event 'cycles', Event count (approx.): 51647128377 \n" - "+ 60.60% 1.33% page_fault2_processes [.] testcase a??\n" - "+ 59.14% 0.41% [kernel] [k] page_fault a??\n" - "+ 58.72% 0.01% [kernel] [k] do_page_fault a??\n" - "+ 58.70% 0.08% [kernel] [k] __do_page_fault a??\n" - "+ 58.50% 0.29% [kernel] [k] handle_mm_fault a??\n" - "+ 40.14% 0.28% [kernel] [k] do_cow_fault a??\n" - "- 34.56% 34.56% [kernel] [k] _raw_spin_lock a??\n" - " - _raw_spin_lock a??\n" - " - 78.11% __res_counter_charge a??\n" - " res_counter_charge a??\n" - " try_charge a??\n" - " - mem_cgroup_try_charge a??\n" - " + 99.99% do_cow_fault a??\n" - " - 10.30% res_counter_uncharge_until a??\n" - " res_counter_uncharge a??\n" - " uncharge_batch a??\n" - " uncharge_list a??\n" - " mem_cgroup_uncharge_list a??\n" - " release_pages a??\n" - " + 4.75% free_pcppages_bulk a??\n" - " + 3.65% do_cow_fault a??\n" - " + 2.24% get_page_from_freelist a??\n" + "+ 60.60% 1.33% page_fault2_processes [.] testcase \342\226\222\n" + "+ 59.14% 0.41% [kernel] [k] page_fault \342\227\206\n" + "+ 58.72% 0.01% [kernel] [k] do_page_fault \342\226\222\n" + "+ 58.70% 0.08% [kernel] [k] __do_page_fault \342\226\222\n" + "+ 58.50% 0.29% [kernel] [k] handle_mm_fault \342\226\222\n" + "+ 40.14% 0.28% [kernel] [k] do_cow_fault \342\226\222\n" + "- 34.56% 34.56% [kernel] [k] _raw_spin_lock \342\226\222\n" + " - _raw_spin_lock \342\226\222\n" + " - 78.11% __res_counter_charge \342\226\222\n" + " res_counter_charge \342\226\222\n" + " try_charge \342\226\222\n" + " - mem_cgroup_try_charge \342\226\222\n" + " + 99.99% do_cow_fault \342\226\222\n" + " - 10.30% res_counter_uncharge_until \342\226\222\n" + " res_counter_uncharge \342\226\222\n" + " uncharge_batch \342\226\222\n" + " uncharge_list \342\226\222\n" + " mem_cgroup_uncharge_list \342\226\222\n" + " release_pages \342\226\222\n" + " + 4.75% free_pcppages_bulk \342\226\222\n" + " + 3.65% do_cow_fault \342\226\222\n" + " + 2.24% get_page_from_freelist \342\226\222\n" "\n" "> You also said that this cost hasn't been there before, but I do see\n" "> that trace in both v3.16 and v3.17-rc3 with roughly the same impact\n" @@ -65,27 +65,21 @@ "\n" "Here's the same workload on the same machine with only Johannes' revert applied:\n" "\n" - "- 35.92% 35.92% [kernel] [k] _raw_spin_lock a??\n" - " - _raw_spin_lock a??\n" - " - 49.09% get_page_from_freelist a??\n" - " - __alloc_pages_nodemask a??\n" - " + 99.90% alloc_pages_vma a??\n" - " - 43.67% free_pcppages_bulk a??\n" - " - 100.00% free_hot_cold_page a??\n" - " + 99.93% free_hot_cold_page_list a??\n" - " - 7.08% do_cow_fault a??\n" - " handle_mm_fault a??\n" - " __do_page_fault a??\n" - " do_page_fault a??\n" - " page_fault a??\n" - " testcase a??\n" + "- 35.92% 35.92% [kernel] [k] _raw_spin_lock \342\226\222\n" + " - _raw_spin_lock \342\226\222\n" + " - 49.09% get_page_from_freelist \342\226\222\n" + " - __alloc_pages_nodemask \342\226\222\n" + " + 99.90% alloc_pages_vma \342\226\222\n" + " - 43.67% free_pcppages_bulk \342\226\222\n" + " - 100.00% free_hot_cold_page \342\226\222\n" + " + 99.93% free_hot_cold_page_list \342\226\222\n" + " - 7.08% do_cow_fault \342\226\222\n" + " handle_mm_fault \342\226\222\n" + " __do_page_fault \342\226\222\n" + " do_page_fault \342\226\222\n" + " page_fault \342\226\222\n" + " testcase \342\226\222\n" "\n" - "So I think it's probably part of the same regression.\n" - "\n" - "--\n" - "To unsubscribe, send a message with 'unsubscribe linux-mm' in\n" - "the body to majordomo@kvack.org. For more info on Linux MM,\n" - "see: http://www.linux-mm.org/ .\n" - "Don't email: <a href=mailto:\"dont@kvack.org\"> email@kvack.org </a>" + So I think it's probably part of the same regression. -efbd436c6775fbfaf2417ce5a89eff3ca69e6d279f19a4888924206db41579ea +b02b444c7694f241ceda74a5559f9a03fe18d3db54fad6551d42332d9648b5c3
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.