Details u have to look at the source code, my guess is based on the following posting:<div><br></div><div><a href="http://kerneltrap.org/mailarchive/linux-kernel/2008/1/23/595569" target="_blank">http://kerneltrap.org/mailarchive/linux-kernel/2008/1/23/595569</a></div>
<div><br></div><div><a href="https://lkml.org/lkml/2011/6/21/249" target="_blank">https://lkml.org/lkml/2011/6/21/249</a></div><div><br></div><div>logic:</div><div><br></div><div>a. u want to be able to do printk from anywhere.</div>
<div>
b. but every call to printk requires a console_sem lock to be acquired. after acquiring console_sem, printk actually serializes the output to a memory buffer. </div><div>c. now problem arises when printk is happening very fast, and so this type of locks is ill-suited for printk().</div>
<div>d. later than this patch is another attempt:</div><div><br></div><div><a href="https://patchwork.kernel.org/patch/1760211/">https://patchwork.kernel.org/patch/1760211/</a></div><div><a href="https://lkml.org/lkml/2012/10/20/90">https://lkml.org/lkml/2012/10/20/90</a></div>
<div><br></div><div>where lazy irq work is being used instead.</div>
<div><br></div><div>read through the comments in the intro to the patch - it covers a lot more than i mentioned here.</div><div><br></div><div>In Documentation/lockdep_design.txt discuss about using irq tracing to trace the lock dependencies.</div>
<div><br></div><div>Lock inversion is a common computer science problem....look up wiki.</div><div><br><div class="gmail_quote">On Sat, Feb 9, 2013 at 10:55 AM, buyitian <span dir="ltr"><<a href="mailto:buyit@live.cn" target="_blank">buyit@live.cn</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div><div dir="ltr"> in the patch 0b5e1c5255e7ee8670e077e8224e5c2281229a5b, it releases console_sem after logbuf_lock, the description of this patch is as below:<br> <br>Release console_sem after unlocking the logbuf_lock so that we don't<br>
generate wakeups while holding logbuf_lock. This avoids some lock<br>inversion troubles once we remove the lockdep_off bits between<br>logbuf_lock and rq->lock (prints while holding rq->lock vs doing<br>wakeups while holding logbuf_lock).<br>
There's of course still an actual deadlock where the printk()s under<br>rq->lock will issue a wakeup from the up() call, but lockdep won't<br>warn about that since semaphores are not tracked. <br> <br> could you please give me a detail example about the issue it tries to fix? thanks.<br>
</div></div>
<br>_______________________________________________<br>
Kernelnewbies mailing list<br>
<a href="mailto:Kernelnewbies@kernelnewbies.org" target="_blank">Kernelnewbies@kernelnewbies.org</a><br>
<a href="http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies" target="_blank">http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies</a><br>
<br></blockquote></div><br><br clear="all"><div><br></div>-- <br>Regards,<br>Peter Teoh
</div>