On Thu, Apr 24, 2014 at 10:59:45AM +0800, Jiang Liu wrote: > On 2014/4/24 1:46, Luck, Tony wrote: > >>>> 1) Handle CPU hot-addition event > >>>> 1.a) gather platform specific information > >>>> 1.b) associate hot-added CPU with a node > >>>> 1.c) create CPU device > >>>> 2) User online hot-added CPUs through sysfs: > >>>> 2.a) cpu_up() > >>>> 2.b) ->try_online_node() > >>>> 2.c) ->hotadd_new_pgdat() > >>>> 2.d) ->node_set_online() > >>>> > >>>> So between 1.b and 2.c, kmalloc_node(nid) may cause invalid > >>>> memory access without the node_online(nid) check. > >>> > >>> Any why was all this not in the Changelog? > >> > >> Also, do explain what kind of hardware you needed to trigger this. This > >> code has been like this for a good while. > > > > With your proposed fix in place the allocations will succeed - but they > > will be done from other nodes ... and this cpu will have to do a remote > > NUMA access for the rest of time. > > > > It would be better to switch the order above - add the memory first, > > then add the cpus. Is that possible? > Hi Tony, > The BIOS always sends CPU hot-addition events before memory > hot-addition events, so it's hard to change the order. > And we couldn't completely solve this performance penalty because the > affected code tries to allocate memory for all possible > CPUs instead of onlined CPUs.
So the BIOS is fucked, news at 11, one would have hoped Intel would have _some_ say in it, but alas. So how about instead you force memory online when you online the first CPU, screw whatever the BIOS does or does not? -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/