On Wed, 19 Feb 2014 19:46:41 -0800 (PST) David Rientjes <rientjes@xxxxxxxxxx> wrote: > On Wed, 19 Feb 2014, Marcelo Tosatti wrote: > > > We agree that, in the future, we'd like to provide the ability to > > dynamically allocate and free 1GB pages at runtime. > > > > Extending the kernel command line interface is a first step. > > > > Do you have a concrete objection to that first step ? > > > > Yes, my concrete objection is that the command line interface is > unnecessary if you can dynamically allocate and free 1GB pages at runtime > unless memory will be so fragmented that it cannot be done when userspace > is brought up. That is not your use case, thus this support is not Yes it is. The early boot is the most reliable moment to allocate huge pages and we want to take advantage from that. I would understand your position against this series if it was intrusive or if it was changing the code in an undesirable way, but the code it adds is completely self-contained and runs only once at boot. It gives us just one more choice for huge pages allocation at boot. I don't see the huge problem here. > needed. I think Mel also brought up this point. > > There's no "first step" about it, this is unnecessary for your use case if > you can do it at runtime. I'm not sure what's so surprising about this. > > > > You can't specify an interleave behavior with Luiz's command line > > > interface so now we'd have two different interfaces for allocating > > > hugepage sizes depending on whether you're specifying a node or not. > > > It's "hugepagesz=1G hugepages=16" vs "hugepage_node=1:16:1G" (and I'd have > > > to look at previous messages in this thread to see if that means 16 1GB > > > pages on node 1 or 1 1GB pages on node 16.) > > > > What syntax do you prefer and why ? > > > > I'm not sure it's interesting to talk about since this patchset is > unnecessary if you can do it at runtime, but since "hugepagesz=" and > "hugepages=" have existed for many kernel releases, we must maintain > backwards compatibility. Thus, it seems, the easiest addition would have > been "hugepagesnode=" which I've mentioned several times, there's no > reason to implement yet another command line option purely as a shorthand > which hugepage_node=1:2:1G is and in a very cryptic way. > -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>