Re: trivial fix for pmemrange

2014-04-05 Thread Miod Vallat
Applied, thanks for spotting the issue, and sorry for the delay. Miod Index: uvm/uvm_pmemrange.c === RCS file: /cvs/src/sys/uvm/uvm_pmemrange.c,v retrieving revision 1.37 diff -p -u -r1.37 uvm_pmemrange.c ---

Re: trivial fix for pmemrange

2014-02-27 Thread Ariane van der Steldt
On 17/02/14 10:11, Kieran Devlin wrote: the original implementation use identical logic for both ‘high’ ‘low’, which will cause ‘high’ ‘low’ end up at same RB-tree node, instead of an expected interval. and finally, break ‘for’ loop logic luckily all these conditions never meet in practice.

trivial fix for pmemrange

2014-02-17 Thread Kieran Devlin
Index: uvm/uvm_pmemrange.c === RCS file: /cvs/src/sys/uvm/uvm_pmemrange.c,v retrieving revision 1.37 diff -p -u -r1.37 uvm_pmemrange.c --- uvm/uvm_pmemrange.c 6 Feb 2014 16:40:40 - 1.37 +++ uvm/uvm_pmemrange.c 13 Feb 2014

Re: trivial fix for pmemrange

2014-02-17 Thread Mike Larkin
On Mon, Feb 17, 2014 at 05:02:34PM +0800, Kieran Devlin wrote: Probably get a better response if you explained what this diff does and/or fixes... -ml Index: uvm/uvm_pmemrange.c === RCS file: /cvs/src/sys/uvm/uvm_pmemrange.c,v

Re: trivial fix for pmemrange

2014-02-17 Thread Kieran Devlin
the original implementation use identical logic for both ‘high’ ‘low’, which will cause ‘high’ ‘low’ end up at same RB-tree node, instead of an expected interval. and finally, break ‘for’ loop logic luckily all these conditions never meet in practice. On Feb 17, 2014, at 5:14 PM, Mike Larkin