| Message ID | 20140128183457.GA9315@linux.vnet.ibm.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Benjamin Herrenschmidt |
| Headers | show |
On Tue, 28 Jan 2014, Nishanth Aravamudan wrote: > Anton Blanchard found an issue with an LPAR that had no memory in Node > 0. Christoph Lameter recommended, as one possible solution, to use > numa_mem_id() for locality of the nearest memory node-wise. However, > numa_mem_id() [and the other related APIs] are only useful if > CONFIG_HAVE_MEMORYLESS_NODES is set. This is only the case for ia64 > currently, but clearly we can have memoryless nodes on ppc64. Add the > Kconfig option and define it to be the same value as CONFIG_NUMA. Well this is trivial but if you need encouragement: Reviewed-by: Christoph Lameter <cl@linux.com>
On 28.01.2014 [10:34:57 -0800], Nishanth Aravamudan wrote: > Anton Blanchard found an issue with an LPAR that had no memory in Node > 0. Christoph Lameter recommended, as one possible solution, to use > numa_mem_id() for locality of the nearest memory node-wise. However, > numa_mem_id() [and the other related APIs] are only useful if > CONFIG_HAVE_MEMORYLESS_NODES is set. This is only the case for ia64 > currently, but clearly we can have memoryless nodes on ppc64. Add the > Kconfig option and define it to be the same value as CONFIG_NUMA. > > On the LPAR in question, which was very inefficiently using slabs, this > took the slab consumption at boot from roughly 7GB to roughly 4GB. Err, this should have been Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> ! Sorry about that Ben! > --- > Ben, the only question I have wrt this change is if it's appropriate to > change it for all powerpc configs (that have NUMA on)? > > diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig > index 25493a0..bb2d5fe 100644 > --- a/arch/powerpc/Kconfig > +++ b/arch/powerpc/Kconfig > @@ -447,6 +447,9 @@ config NODES_SHIFT > default "4" > depends on NEED_MULTIPLE_NODES > > +config HAVE_MEMORYLESS_NODES > + def_bool NUMA > + > config ARCH_SELECT_MEMORY_MODEL > def_bool y > depends on PPC64
On Thu, 13 Feb 2014, Nishanth Aravamudan wrote: > > Anton Blanchard found an issue with an LPAR that had no memory in Node > > 0. Christoph Lameter recommended, as one possible solution, to use > > numa_mem_id() for locality of the nearest memory node-wise. However, > > numa_mem_id() [and the other related APIs] are only useful if > > CONFIG_HAVE_MEMORYLESS_NODES is set. This is only the case for ia64 > > currently, but clearly we can have memoryless nodes on ppc64. Add the > > Kconfig option and define it to be the same value as CONFIG_NUMA. > > > > On the LPAR in question, which was very inefficiently using slabs, this > > took the slab consumption at boot from roughly 7GB to roughly 4GB. > > Err, this should have been > > Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> > > ! > > Sorry about that Ben! > > > --- > > Ben, the only question I have wrt this change is if it's appropriate to > > change it for all powerpc configs (that have NUMA on)? > > I'm suspecting that Ben will request that the proper set_numa_mem() calls are done for ppc init to make this actually do anything other than return numa_mem_id() == numa_node_id(). > > diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig > > index 25493a0..bb2d5fe 100644 > > --- a/arch/powerpc/Kconfig > > +++ b/arch/powerpc/Kconfig > > @@ -447,6 +447,9 @@ config NODES_SHIFT > > default "4" > > depends on NEED_MULTIPLE_NODES > > > > +config HAVE_MEMORYLESS_NODES > > + def_bool NUMA > > + > > config ARCH_SELECT_MEMORY_MODEL > > def_bool y > > depends on PPC64
On 13.02.2014 [14:45:49 -0800], David Rientjes wrote: > On Thu, 13 Feb 2014, Nishanth Aravamudan wrote: > > > > Anton Blanchard found an issue with an LPAR that had no memory in Node > > > 0. Christoph Lameter recommended, as one possible solution, to use > > > numa_mem_id() for locality of the nearest memory node-wise. However, > > > numa_mem_id() [and the other related APIs] are only useful if > > > CONFIG_HAVE_MEMORYLESS_NODES is set. This is only the case for ia64 > > > currently, but clearly we can have memoryless nodes on ppc64. Add the > > > Kconfig option and define it to be the same value as CONFIG_NUMA. > > > > > > On the LPAR in question, which was very inefficiently using slabs, this > > > took the slab consumption at boot from roughly 7GB to roughly 4GB. > > > > Err, this should have been > > > > Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> > > > > ! > > > > Sorry about that Ben! > > > > > --- > > > Ben, the only question I have wrt this change is if it's appropriate to > > > change it for all powerpc configs (that have NUMA on)? > > > > > I'm suspecting that Ben will request that the proper set_numa_mem() calls > are done for ppc init to make this actually do anything other than return > numa_mem_id() == numa_node_id(). You're right, thanks for pointing this out. I could have sworn that in my previous debugging I saw proper NUMA information, but perhaps it was just correct based upon the system configuration. > > > diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig > > > index 25493a0..bb2d5fe 100644 > > > --- a/arch/powerpc/Kconfig > > > +++ b/arch/powerpc/Kconfig > > > @@ -447,6 +447,9 @@ config NODES_SHIFT > > > default "4" > > > depends on NEED_MULTIPLE_NODES > > > > > > +config HAVE_MEMORYLESS_NODES > > > + def_bool NUMA > > > + > > > config ARCH_SELECT_MEMORY_MODEL > > > def_bool y > > > depends on PPC64 >
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 25493a0..bb2d5fe 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -447,6 +447,9 @@ config NODES_SHIFT default "4" depends on NEED_MULTIPLE_NODES +config HAVE_MEMORYLESS_NODES + def_bool NUMA + config ARCH_SELECT_MEMORY_MODEL def_bool y depends on PPC64
Anton Blanchard found an issue with an LPAR that had no memory in Node 0. Christoph Lameter recommended, as one possible solution, to use numa_mem_id() for locality of the nearest memory node-wise. However, numa_mem_id() [and the other related APIs] are only useful if CONFIG_HAVE_MEMORYLESS_NODES is set. This is only the case for ia64 currently, but clearly we can have memoryless nodes on ppc64. Add the Kconfig option and define it to be the same value as CONFIG_NUMA. On the LPAR in question, which was very inefficiently using slabs, this took the slab consumption at boot from roughly 7GB to roughly 4GB. --- Ben, the only question I have wrt this change is if it's appropriate to change it for all powerpc configs (that have NUMA on)?