tag:blogger.com,1999:blog-6469704299235308349.post5472708000195561029..comments2024-02-09T16:24:46.087+11:00Comments on -ck hacking: 3.7-ck1, BFS 426 for linux-3.7ckhttp://www.blogger.com/profile/02904761195451530213noreply@blogger.comBlogger74125tag:blogger.com,1999:blog-6469704299235308349.post-79255126358342108412014-12-23T09:54:26.516+11:002014-12-23T09:54:26.516+11:00Interesting thoughts, Con. Hope people take note.....Interesting thoughts, Con. Hope people take note...Jos Poortvliethttps://www.blogger.com/profile/05243886270488333877noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-59125838483501676702013-05-17T21:45:32.679+10:002013-05-17T21:45:32.679+10:00I tend to find BFS doesn't play nice with nice...I tend to find BFS doesn't play nice with nice. When I run a processor-intensive task such as a backup or compile in the background niced so I can keep working on a responsive machine, BFS seems to thrash, pushing the load up, grinding the desktop to a halt, and taking around 3 times as long to complete the task as the normal kernel scheduler.<br /><br />For this reason I have always avoided BFS-enabled kernels, but the distro I use (PCLinuxOS) has now removed the non-BFS kernels from its repository, so it looks as if I will be faced with a choice between BFS kernels or compiling my own in future.<br /><br />One of the kernel packagers has suggested I report the problem here.<br />kjpetrienoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-20382531483350795712013-02-11T22:05:59.154+11:002013-02-11T22:05:59.154+11:00Darn :\Darn :\ckhttps://www.blogger.com/profile/02904761195451530213noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-4290061295829385682013-02-11T22:02:24.183+11:002013-02-11T22:02:24.183+11:00yes, confirmed, i have other one today and have be...yes, confirmed, i have other one today and have been with the normal desktop use, no using rsnapshot. http://pastebin.com/dQMyHp3emraserohttps://www.blogger.com/profile/02887878973092924351noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-10324732398422276192013-02-04T01:59:12.504+11:002013-02-04T01:59:12.504+11:00it seems to be failing still, here you have the la...it seems to be failing still, here you have the last error http://pastebin.com/Axb62q0Zmraserohttps://www.blogger.com/profile/02887878973092924351noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-73014146680943003712013-01-31T02:29:54.706+11:002013-01-31T02:29:54.706+11:00I have tried two consecutive backups and the probl...I have tried two consecutive backups and the problem didn't show up, i will be testing the next few days but seems to be fixed or more difficult to trigger.<br /><br />On the other side i have a new kernel problem but i don't know if maybe can be a VirtualBox driver problem, i pasted here http://pastebin.com/yAKVWk9Fmraserohttps://www.blogger.com/profile/02887878973092924351noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-1612847042678724992013-01-30T21:51:34.779+11:002013-01-30T21:51:34.779+11:00i will test ASAP and post here the results, thanks...i will test ASAP and post here the results, thanks!mraserohttps://www.blogger.com/profile/02887878973092924351noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-9031632964292131692013-01-30T21:47:35.000+11:002013-01-30T21:47:35.000+11:00Would be interesting to see if BFS 427 has any eff...Would be interesting to see if BFS 427 has any effect on this bug of yours.ckhttps://www.blogger.com/profile/02904761195451530213noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-9466185915287128222013-01-29T15:27:21.781+11:002013-01-29T15:27:21.781+11:00The accounting for interrupt time becomes less acc...The accounting for interrupt time becomes less accurate. The code overhead is less but I've been unable to demonstrate a performance difference.ckhttps://www.blogger.com/profile/02904761195451530213noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-84877705272860069272013-01-29T15:22:57.839+11:002013-01-29T15:22:57.839+11:00Thank you for the quick reply! :-)
Just to be sur...Thank you for the quick reply! :-)<br /><br />Just to be sure: Does _not_ setting IRQ_TIME_ACCOUNTING but TICK_CPU_ACCOUNTING = y manually do something harmful or anything better?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-77734372321971168512013-01-29T14:56:36.997+11:002013-01-29T14:56:36.997+11:00It's a very interesting topic. CK should have ...It's a very interesting topic. CK should have a look at it. As I remember, BFS now is kind of HT awared, in task switching, it will stick to the cache-hot core/virtual core.Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-39379526511701226362013-01-29T12:30:24.698+11:002013-01-29T12:30:24.698+11:00The CPU accounting in BFS by default is already us...The CPU accounting in BFS by default is already using high resolution just like the mainline "IRQ TIME ACCOUNTING" so it actually doesn't change anything. It's just the new visible kernel option in mainline. I've investigated and enabling both does nothing harmful to BFS.ckhttps://www.blogger.com/profile/02904761195451530213noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-91533706610066481962013-01-29T12:14:30.139+11:002013-01-29T12:14:30.139+11:00This was introduced with BFS 426 for linux-3.7 (th...This was introduced with BFS 426 for linux-3.7 (the hunk @ line 754) and is not present in the previous versions.<br /><br />Is there a reason to prefer IRQ_TIME_ACCOUNTING over TICK_CPU_ACCOUNTING? ( @ck: do I read the patch correctly that this is what you intended? ) The config help says, "... so there can be a<br />small performance impact" with it. And, ck, do you know what takes precedence when wrongly having both set to y or if this has side effects?<br /><br />Thank you, Manuel Krause<br /><br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-59687904418471781372013-01-29T08:40:33.532+11:002013-01-29T08:40:33.532+11:00ck, thank you for your continued contributions.ck, thank you for your continued contributions.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-72544876092489011682013-01-28T20:50:00.308+11:002013-01-28T20:50:00.308+11:00I have to say thanks, for bfs and all the work and...I have to say thanks, for bfs and all the work and effort you invest in it. <br /><br />All the best for you and your beloved.<br /><br />GoldfieldAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-52858696972865986462013-01-28T09:30:11.437+11:002013-01-28T09:30:11.437+11:00That could be a very relevant finding with respect...That could be a very relevant finding with respect to bugs, thanks for pointing it out!ckhttps://www.blogger.com/profile/02904761195451530213noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-58837129503613466602013-01-28T03:47:18.107+11:002013-01-28T03:47:18.107+11:00if someone is interested....
3.7.4 cfs
CONFIG_NO_...if someone is interested....<br /><br />3.7.4 cfs<br />CONFIG_NO_HZ=y<br />real 8m18.377s<br />user 30m47.102s<br />sys 0m15.832s<br /><br />3.7.4 bfs<br />CONFIG_NO_HZ=y<br />real 8m49.038s<br />user 32m30.408s<br />sys 0m4.049s<br /><br />3.7.4 bfs<br /># CONFIG_NO_HZ is not set<br />real 9m40.055s<br />user 36m15.898s<br />sys 0m5.635sAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-29164918469479863192013-01-28T03:34:55.844+11:002013-01-28T03:34:55.844+11:00normaly i can choose between TICK_CPU_ACCOUNTING o...normaly i can choose between TICK_CPU_ACCOUNTING or IRQ_TIME_ACCOUNTING<br /><br />With bfs, both! will be enabled in .config<br /><br />CONFIG_TICK_CPU_ACCOUNTING=y<br />CONFIG_IRQ_TIME_ACCOUNTING=y<br /><br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-54527655438545617022013-01-26T19:20:43.732+11:002013-01-26T19:20:43.732+11:00Thank you for make my desktop usable :), i can tes...Thank you for make my desktop usable :), i can test any patch you send me or any update you do in bfs if needed, it's very easy to me to trigger this bug.mraserohttps://www.blogger.com/profile/02887878973092924351noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-8347592458615141222013-01-26T09:47:34.451+11:002013-01-26T09:47:34.451+11:00That could well be. There is a bug striking in wei...That could well be. There is a bug striking in weird and wonderful places and this might be the same bug. I'm investigating, thanks.ckhttps://www.blogger.com/profile/02904761195451530213noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-73509383261661820382013-01-26T09:46:57.182+11:002013-01-26T09:46:57.182+11:00As I already said in my previous response, it does...As I already said in my previous response, it does move to unused cores before thread idle siblings on busy cores.ckhttps://www.blogger.com/profile/02904761195451530213noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-10440381725236765352013-01-26T04:19:26.300+11:002013-01-26T04:19:26.300+11:00I am having in the last versions, from 3.6.x upwar...I am having in the last versions, from 3.6.x upwards i think, problems making backups with rsnapshot (a rsync based backup solution) to my mdadm software raid 5 xfs filesystem, here you can see a thread where i reported the problem to xfs mailing list, but seems to be related to the kernel, could this be a bfs problem?<br /><br />http://oss.sgi.com/archives/xfs/2013-01/msg00081.html<br />mraserohttps://www.blogger.com/profile/02887878973092924351noreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-36755517725723184302013-01-26T03:19:06.899+11:002013-01-26T03:19:06.899+11:00Here is a lengthy article of what I actually meant...Here is a lengthy article of what I actually meant with "real" cores.<br />http://software.intel.com/en-us/articles/performance-insights-to-intel-hyper-threading-technologyAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-71325614986241605222013-01-24T19:57:19.593+11:002013-01-24T19:57:19.593+11:00Thanks for clarifying!Thanks for clarifying!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-6469704299235308349.post-74763320144841134702013-01-24T19:30:15.274+11:002013-01-24T19:30:15.274+11:00There is no such thing as "real" cores v...There is no such thing as "real" cores versus siblings. Siblings only become SMT siblings when something is bound to the other thread unit on a core. Yes, BFS already does bind to unused "cores" before trying siblings of busy units. However, if they're all in use, it will then find a sibling in the interests of latency rather than hold off and wait to get back on the same core.ckhttps://www.blogger.com/profile/02904761195451530213noreply@blogger.com