Re: [PATCH] [REPOST #2] mm: show node to memory sectionrelationship with symlinks in sysfs

[Gary Hade]

On Fri, Nov 14, 2008 at 08:41:25AM -0800, Gary Hade wrote:
> On Fri, Nov 14, 2008 at 08:05:17AM -0800, Badari Pulavarty wrote:
> > On Thu, 2008-11-13 at 08:54 -0800, Gary Hade wrote:
> > > On Wed, Nov 12, 2008 at 02:16:15PM -0800, Badari Pulavarty wrote:
> > > > On Mon, 2008-11-03 at 15:48 -0800, Gary Hade wrote:
> > > > > Show node to memory section relationship with symlinks in sysfs
> > > > > 
> > > > > Add /sys/devices/system/node/nodeX/memoryY symlinks for all
> > > > > the memory sections located on nodeX.  For example:
> > > > > /sys/devices/system/node/node1/memory135 -> ../../memory/memory135
> > > > > indicates that memory section 135 resides on node1.
> > > > > 
> > > > > Also revises documentation to cover this change as well as updating
> > > > > Documentation/ABI/testing/sysfs-devices-memory to include descriptions
> > > > > of memory hotremove files 'phys_device', 'phys_index', and 'state'
> > > > > that were previously not described there.
> > > > > 
> > > > > In addition to it always being a good policy to provide users with
> > > > > the maximum possible amount of physical location information for
> > > > > resources that can be hot-added and/or hot-removed, the following
> > > > > are some (but likely not all) of the user benefits provided by
> > > > > this change.
> > > > > Immediate:
> > > > >   - Provides information needed to determine the specific node
> > > > >     on which a defective DIMM is located.  This will reduce system
> > > > >     downtime when the node or defective DIMM is swapped out.
> > > > >   - Prevents unintended onlining of a memory section that was 
> > > > >     previously offlined due to a defective DIMM.  This could happen
> > > > >     during node hot-add when the user or node hot-add assist script
> > > > >     onlines _all_ offlined sections due to user or script inability
> > > > >     to identify the specific memory sections located on the hot-added
> > > > >     node.  The consequences of reintroducing the defective memory
> > > > >     could be ugly.
> > > > >   - Provides information needed to vary the amount and distribution
> > > > >     of memory on specific nodes for testing or debugging purposes.
> > > > > Future:
> > > > >   - Will provide information needed to identify the memory
> > > > >     sections that need to be offlined prior to physical removal
> > > > >     of a specific node.
> > > > > 
> > > > > Symlink creation during boot was tested on 2-node x86_64, 2-node
> > > > > ppc64, and 2-node ia64 systems.  Symlink creation during physical
> > > > > memory hot-add tested on a 2-node x86_64 system.
> > > > > 
> > > > > Supersedes the "mm: show memory section to node relationship in sysfs"
> > > > > patch posted on 05 Sept 2008 which created node ID containing 'node'
> > > > > files in /sys/devices/system/memory/memoryX instead of symlinks.
> > > > > Changed from files to symlinks due to feedback that symlinks were
> > > > > more consistent with the sysfs way.
> > > > > 
> > > > > Supersedes the "mm: show node to memory section relationship with
> > > > > symlinks in sysfs" patch posted on 29 Sept 2008 to address a Yasunori
> > > > > Goto reported problem where an incorrect symlink was created due to
> > > > > a range of uninitialized pages at the beginning of a section.  This
> > > > > problem which produced a symlink in /sys/devices/system/node/node0
> > > > > that incorrectly referenced a mem section located on node1 is corrected
> > > > > in this version.  This version also covers the case were a mem section
> > > > > could span multiple nodes.
> > > > > 
> > > > > Supersedes the "mm: show node to memory section relationship with
> > > > > symlinks in sysfs" patch posted on 09 Oct 2008 to add the Andrew
> > > > > Morton requested usefulness information and update to apply cleanly
> > > > > to 2.6.28-rc3 and 2.6-git.  Code is unchanged.
> > > > > 
> > > > > Signed-off-by: Gary Hade <garyhade@xxxxxxxxxx>
> > > > > Signed-off-by: Badari Pulavarty <pbadari@xxxxxxxxxx>
> > > > > 
> > > > 
> > > > Hi Gary,
> > > > 
> > > > While testing latest mmtom (which has this patch) ran into an issue
> > > > with sysfs files. What I noticed was, with this patch "memoryXX"
> > > > directories in /sys/devices/system/memory/ are not getting cleaned up.
> > > > Backing out the patch seems to fix the problem. 
> > > > 
> > > > When I tried to remove 64 blocks of memory, empty  directories are
> > > > stayed around. (look at memory151 - memory215). This is causing OOPS
> > > > while trying to add memory block again. I think this could be because 
> > > > of the symlink added from node directory.  Can you look ?
> > > 
> > > Badari, The call to unregister_mem_sect_under_nodes() in
> > > remove_memory_block() preceding the removal of the files in
> > > the memory section directory _should have_ removed all the
> > > symlinks referencing the memory section directory.  Did you
> > > happen to check to see if the symlinks to memory151-memory215
> > > were still present?
> > > 
> > > Gary
> > > 
> > 
> > Hi Gary,
> > 
> > As discussed earlier, patch is leaving an extra reference on the
> > memoryX directory. Needs a kobject_put() to match the reference
> > you get in find_memory_block().
> 
> Badari, Thanks again for finding that!
> 
> > 
> > Could you update the patch and resend it ?
> 
> Will do.

I just posted a replacement patch with subject line
  [PATCH] [REPOST #3] mm: show node to memory section relationship
                      with symlinks in sysfs
In addition to addressing the memory section directory removal problem
it also contains a change to correct a 'for' loop early termination
problem that I noticed while debugging the directory removal problem.

Andrew, If it would be better from a -mm standpoint for me to give
you a small patch that would apply on top of instead of replacing
the bad one, just let me know.

Thanks,
Gary

-- 
Gary Hade
System x Enablement
IBM Linux Technology Center
503-578-4503  IBM T/L: 775-4503
garyhade@xxxxxxxxxx
http://www.ibm.com/linux/ltc

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/