diff for duplicates of <96341387285682@web18g.yandex.ru> diff --git a/a/1.txt b/N1/1.txt index 91ab44a..a9508d3 100644 --- a/a/1.txt +++ b/N1/1.txt @@ -1,27 +1,27 @@ 17.12.2013, 16:47, "Peter Zijlstra" <peterz@infradead.org>: > On Tue, Dec 17, 2013 at 04:02:58PM +0400, Kirill Tkhai wrote: > ->> 13.12.2013, 19:42, "Peter Zijlstra" <peterz@infradead.org>: ->>> On Wed, Nov 27, 2013 at 07:59:13PM +0400, Kirill Tkhai wrote: ->>>> This patch touches RT group scheduling case. +>> О©╫13.12.2013, 19:42, "Peter Zijlstra" <peterz@infradead.org>: +>>> О©╫On Wed, Nov 27, 2013 at 07:59:13PM +0400, Kirill Tkhai wrote: +>>>> О©╫О©╫This patch touches RT group scheduling case. >>>> ->>>> Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority, ->>>> while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because ->>>> changing of priority on a child level does not guarantee that the priority is ->>>> the highest all over the rq. So, this leak makes RT balancing unusable. +>>>> О©╫О©╫Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority, +>>>> О©╫О©╫while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because +>>>> О©╫О©╫changing of priority on a child level does not guarantee that the priority is +>>>> О©╫О©╫the highest all over the rq. So, this leak makes RT balancing unusable. >>>> ->>>> The short example: the task having the highest priority among all rq's RT tasks ->>>> (no one other task has the same priority) are waking on a throttle rt_rq. ->>>> The rq's cpupri is set to the task's priority equivalent, but real ->>>> rq->rt.highest_prio.curr is less. +>>>> О©╫О©╫The short example: the task having the highest priority among all rq's RT tasks +>>>> О©╫О©╫(no one other task has the same priority) are waking on a throttle rt_rq. +>>>> О©╫О©╫The rq's cpupri is set to the task's priority equivalent, but real +>>>> О©╫О©╫rq->rt.highest_prio.curr is less. >>>> ->>>> The patch below fixes the problem. +>>>> О©╫О©╫The patch below fixes the problem. >>>> ->>>> It looks like all version have this bug, so I CC'ed stable mailing list. ->>> Yeah, I think this is right. +>>>> О©╫О©╫It looks like all version have this bug, so I CC'ed stable mailing list. +>>> О©╫Yeah, I think this is right. >>> ->>> cpupri stuff should indeed only be changed for the top level group. ->> Ingo, are you going to apply this patch? Or will you give any comments? +>>> О©╫cpupri stuff should indeed only be changed for the top level group. +>> О©╫Ingo, are you going to apply this patch? Or will you give any comments? > > I queued it, Ingo should get it through me somewhere today if all things > go well. diff --git a/a/content_digest b/N1/content_digest index 0f95d38..c8c637e 100644 --- a/a/content_digest +++ b/N1/content_digest @@ -15,27 +15,27 @@ "17.12.2013, 16:47, \"Peter Zijlstra\" <peterz@infradead.org>:\n" "> On Tue, Dec 17, 2013 at 04:02:58PM +0400, Kirill Tkhai wrote:\n" ">\n" - ">> \302\24013.12.2013, 19:42, \"Peter Zijlstra\" <peterz@infradead.org>:\n" - ">>> \302\240On Wed, Nov 27, 2013 at 07:59:13PM +0400, Kirill Tkhai wrote:\n" - ">>>> \302\240\302\240This patch touches RT group scheduling case.\n" + ">> \320\236\302\251\342\225\25313.12.2013, 19:42, \"Peter Zijlstra\" <peterz@infradead.org>:\n" + ">>> \320\236\302\251\342\225\253On Wed, Nov 27, 2013 at 07:59:13PM +0400, Kirill Tkhai wrote:\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253This patch touches RT group scheduling case.\n" ">>>>\n" - ">>>> \302\240\302\240Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority,\n" - ">>>> \302\240\302\240while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because\n" - ">>>> \302\240\302\240changing of priority on a child level does not guarantee that the priority is\n" - ">>>> \302\240\302\240the highest all over the rq. So, this leak makes RT balancing unusable.\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253Functions inc_rt_prio_smp() and dec_rt_prio_smp() change (global) rq's priority,\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253while rt_rq passed to them may be not the top-level rt_rq. This is wrong, because\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253changing of priority on a child level does not guarantee that the priority is\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253the highest all over the rq. So, this leak makes RT balancing unusable.\n" ">>>>\n" - ">>>> \302\240\302\240The short example: the task having the highest priority among all rq's RT tasks\n" - ">>>> \302\240\302\240(no one other task has the same priority) are waking on a throttle rt_rq.\n" - ">>>> \302\240\302\240The rq's cpupri is set to the task's priority equivalent, but real\n" - ">>>> \302\240\302\240rq->rt.highest_prio.curr is less.\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253The short example: the task having the highest priority among all rq's RT tasks\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253(no one other task has the same priority) are waking on a throttle rt_rq.\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253The rq's cpupri is set to the task's priority equivalent, but real\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253rq->rt.highest_prio.curr is less.\n" ">>>>\n" - ">>>> \302\240\302\240The patch below fixes the problem.\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253The patch below fixes the problem.\n" ">>>>\n" - ">>>> \302\240\302\240It looks like all version have this bug, so I CC'ed stable mailing list.\n" - ">>> \302\240Yeah, I think this is right.\n" + ">>>> \320\236\302\251\342\225\253\320\236\302\251\342\225\253It looks like all version have this bug, so I CC'ed stable mailing list.\n" + ">>> \320\236\302\251\342\225\253Yeah, I think this is right.\n" ">>>\n" - ">>> \302\240cpupri stuff should indeed only be changed for the top level group.\n" - ">> \302\240Ingo, are you going to apply this patch? Or will you give any comments?\n" + ">>> \320\236\302\251\342\225\253cpupri stuff should indeed only be changed for the top level group.\n" + ">> \320\236\302\251\342\225\253Ingo, are you going to apply this patch? Or will you give any comments?\n" ">\n" "> I queued it, Ingo should get it through me somewhere today if all things\n" "> go well.\n" @@ -44,4 +44,4 @@ "\n" Thanks, Peter -f9115f6bed3f59a7c36508f89f975c91a7ff4ca09ceb9504e53f92faf86507e7 +cf0d78b446c21b3bb9c306e9b80bd0947c6eac2dfb14c2919299fc994c8812eb
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.