Subject : 2.6.33 pagemap endless read loop Submitter : Andi Kleen <andi@firstfloor.org> Date : 2010-01-10 2:09 References : http://marc.info/?l=linux-kernel&m=126308941423848&w=4 Notify-Also : Américo Wang <xiyou.wangcong@gmail.com> This entry is being used for tracking a regression from 2.6.32. Please don't close it until the problem is fixed in the mainline.
> 2.6.33 pagemap endless read loop LTP bug proc01 wants info about every page in virtual address space on x86_64.
On Thursday 14 January 2010, Matt Mackall wrote: > On Thu, 2010-01-14 at 00:50 +0100, Andi Kleen wrote: > > On Wed, Jan 13, 2010 at 10:57:34PM +0100, Rafael J. Wysocki wrote: > > > On Wednesday 13 January 2010, Américo Wang wrote: > > > > On Mon, Jan 11, 2010 at 6:32 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote: > > > > > This message has been generated automatically as a part of a report > > > > > of recent regressions. > > > > > > > > > > The following bug entry is on the current list of known regressions > > > > > from 2.6.32. Please verify if it still should be listed and let me > know > > > > > (either way). > > > > > > > > > > > > > > > Bug-Entry : http://bugzilla.kernel.org/show_bug.cgi?id=15041 > > > > > Subject : Pagemap endless read loop with LTP > > > > > Submitter : Andi Kleen <andi@firstfloor.org> > > > > > Date : 2010-01-10 2:09 (1 days old) > > > > > References : > http://marc.info/?l=linux-kernel&m=126308941423848&w=4 > > > > > > > > > > > > > According to Andi's later reply, it shoud not be an endless loop, > > > > it just takes a rather long time. > > > > > > Andi? > > > > It'll try to read 47 bits worth of 0s on 64bit x86-64. > > > > I haven't tried to wait until that finishes, but I estimate a few months of > > CPU time at least. > > > > The interface is just misdesigned, but I guess we cannot do > > anything about that for now. > > It's perfectly sensible. What's not sensible is reading the entirety of > everything you find in /proc, something that just about every Linux > admin figures out moments after running their first recursive grep. > > > LTP will just need to do a workaround. > > Or they could actually, you know, add a test of pagemap. Not much chance > of that, though - CVS reports it took them until 2005 to figure out that > skipping /proc/kcore was a good idea.