tag:blogger.com,1999:blog-2963790426029213933.post982484753013808672..comments2024-02-29T00:33:07.382-08:00Comments on Alfred Chen's Blog: BFS0461 linux 4.0 syncup patchAlfred Chenhttp://www.blogger.com/profile/03164306846702841944noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-2963790426029213933.post-52169449501962216302015-04-15T17:52:21.825-07:002015-04-15T17:52:21.825-07:00This is a minor change from main stream. In real w...This is a minor change from main stream. In real world, it does "nothing", :)<br />But from the coder's view, I'd like the updated version, it's easy to understand and "may be" a few final instructions saved. But consider the chances sched_show_task() being called, it does "nothing".Alfred Chenhttps://www.blogger.com/profile/03164306846702841944noreply@blogger.comtag:blogger.com,1999:blog-2963790426029213933.post-40260163181921044282015-04-15T13:25:14.387-07:002015-04-15T13:25:14.387-07:00Hi, Alfred!
Many thanks for this sync-up! I haven&...Hi, Alfred!<br />Many thanks for this sync-up! I haven't tried kernel 4.0.0 yet, as there are too many other patches outstanding, but have added some of the chunks from your patch to my 3.19(.4) setup. These chunks add the missing kernel/sched/core.c code that I wasn't able to safely adapt to bfs.c on my own at the times when kernelOfTruth reported the row of cfs improvement/fix-patches @ ck's blog.<br />This 3.19.4, together with your -gc patches, bfq and tuxonice has only few uptime hours so far, but I'm glad not to see any issues.<br /><br />Just out of curiosity: What does this chunk change "in real world"?:<br />> @@ -4944,9 +4947,10 @@ void sched_show_task(struct task_struct *p)<br />> {<br />> unsigned long free = 0;<br />> int ppid;<br />> - unsigned state;<br />> + unsigned long state = p->state;<br />> <br />> - state = p->state ? __ffs(p->state) + 1 : 0;<br />> + if (state)<br />> + state = __ffs(state) + 1;<br />> printk(KERN_INFO "%-15.15s %c", p->comm,<br />> state < sizeof(stat_nam) - 1 ? stat_nam[state] : '?');<br />> #if BITS_PER_LONG == 32<br /><br />BR ManuelAnonymousnoreply@blogger.com