Hello Shrikanth, On 10/7/2026 8:13 PM, Shrikanth Hegde wrote: > Hi Prateek. > > On 10/2/26 12:58 AM, K Prateek Nayak wrote: >> Initialize sparsebitmap (sbm) topology based on the coregroup >> information. Each coregorup gets its own sparsemask leaf. >> >> pSeries and memory hotplug are interesting since a hotplug can place a >> newly added CPU on any online node. This requires special care to allow >> estimating bitmask size considering the worst case scenarios - each CPU >> onlined is on a separate node, and this is a new N_CPU node. >> >> Platform may enforce a stricter standards for the CPUs being online and >> what nodes they can be mapped to but the current implementations makes >> no assumptions and considers each CPU can be onlined on a unique node. >> > > I think this suffers the same fate as structures which are allocated at boot > time such as runqueues. > > So your fallback option of putting all the disabled into singleton node may be > sensible option. (If you are not doing that already)
For pSeries I allocate a humongous chunk of memory on sbm_alloc() and control how much of that memory needs traversing based on the set of online CPUs. Not ideal but yes, I will probably pivot to doing something like what per-cpu setup does by using ealy topology + assuming a few things about the rest. > But yhea, will see more into it, this changing node stuff is new for me too. > Also i need to read your patch series too :) Thank you again for taking a look at the series and early testing. _Me goes digs down on couple regressions I've noticed_ -- Thanks and Regards, Prateek
