Messages in this thread | | | Subject | Re: [2.6.24-rc8-mm1][regression?] numactl --interleave=all doesn't works on memoryless node. | From | Lee Schermerhorn <> | Date | Tue, 05 Feb 2008 17:04:30 -0500 |
| |
On Tue, 2008-02-05 at 15:33 -0600, Paul Jackson wrote: > David wrote: > > It would be disappointing to see a lot of work done to fix > > The suggested patch of KOSAKI Motohiro didn't look like a lot of work to me. > > I continue to prefer not to hijack this thread for that other discussion. > Just presenting your position and calling it "simple" is misleading. > The discussion so far has involved over a hundred messages over months, > and certainly your position, nor mine for that matter, obtained concensus. > > How does the patch of KOSAKI Motohiro, earlier in this thread, look to you? >
Paul,
It wasn't clear to me whether Kosaki-san's patch required a modified numactl/libnuma or not. I think so, because that patch doesn't change the error return in contextualize_policy() and in mpol_check_policy(). My modified numactl/libnuma avoids this by only passing in allowed mems fetch via get_mempolicy() with the new MEMS_ALLOWED flags.
The patch I just posted doesn't depend on the numactl changes and seems quite minimal to me. I think it cleans up the differences between set_mempolicy() and mbind(), as well. However, some may take exception to the change in behavior--silently ignoring dis-allowed nodes in set_mempolicy().
Also, your cpuset/mempolicy work will probably need to undo the unconditional masking in contextualize_policy() and/or save the original node mask somewhere...
Lee
| |