On 5/6/21 12:14 AM, Dan Williams wrote: > On Wed, May 5, 2021 at 3:38 PM Joao Martins <joao.m.martins@xxxxxxxxxx> wrote: >> >> >> >> On 5/5/21 11:34 PM, Dan Williams wrote: >>> On Thu, Mar 25, 2021 at 4:10 PM Joao Martins <joao.m.martins@xxxxxxxxxx> wrote: >>>> >>>> @altmap is stored in a dev_pagemap, but it will be repurposed for >>>> hotplug memory for storing the memmap in the hotplugged memory[*] and >>>> reusing the altmap infrastructure to that end. This is to say that >>>> altmap can't be replaced with a @pgmap as it is going to cover more than >>>> dev_pagemap backend altmaps. >>> >>> I was going to say, just pass the pgmap and lookup the altmap from >>> pgmap, but Oscar added a use case for altmap independent of pgmap. So >>> you might refresh this commit message to clarify why passing pgmap by >>> itself is not sufficient. >>> >> Isn't that what I am doing above with that exact paragraph? I even >> reference his series at the end of commit description :) in [*] > > Oh, sorry, it didn't hit me explicitly that you were talking about > Oscar's work I thought you were referring to your own changes in this > set. I see it now... at a minimum the tense needs updating since > Oscar's changes are in the past not the future anymore. /me nods > If it helps, > the following reads more direct to me: "In support of using compound > pages for devmap mappings, plumb the pgmap down to the > vmemmap_populate* implementation. Note that while altmap is > retrievable from pgmap the memory hotplug codes passes altmap without > pgmap, so both need to be independently plumbed." > Let me use this one instead, definitely better than my strange english :)
