* Memory hotplug vmem pages
@ 2018-10-12 8:15 Zaslonko Mikhail
2018-10-30 13:44 ` Michal Hocko
0 siblings, 1 reply; 3+ messages in thread
From: Zaslonko Mikhail @ 2018-10-12 8:15 UTC (permalink / raw)
To: Michal Hocko; +Cc: linux-mm
Hello Michal,
I've read a recent discussion about introducing the memory types for
memory hotplug:
https://marc.info/?t=153814716600004&r=1&w=2
In particular I was interested in the idea of moving vmem struct pages
to the hotplugable memory itself. I'm also looking into it for s390
right now. So, in one of your replies you mentioned that you "have
proposed (but haven't finished this due to other stuff) a solution for
this". Have you covered any part of that solution yet? Could you please
point me to any relevant discussions on this matter?
Thanks,
Mikhail Zaslonko
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Memory hotplug vmem pages
2018-10-12 8:15 Memory hotplug vmem pages Zaslonko Mikhail
@ 2018-10-30 13:44 ` Michal Hocko
2018-10-31 15:52 ` Zaslonko Mikhail
0 siblings, 1 reply; 3+ messages in thread
From: Michal Hocko @ 2018-10-30 13:44 UTC (permalink / raw)
To: Zaslonko Mikhail; +Cc: linux-mm
[Sorry for late response]
On Fri 12-10-18 10:15:26, Zaslonko Mikhail wrote:
> Hello Michal,
>
> I've read a recent discussion about introducing the memory types for memory
> hotplug:
> https://marc.info/?t=153814716600004&r=1&w=2
>
> In particular I was interested in the idea of moving vmem struct pages to
> the hotplugable memory itself. I'm also looking into it for s390 right now.
> So, in one of your replies you mentioned that you "have proposed (but
> haven't finished this due to other stuff) a solution for this". Have you
> covered any part of that solution yet? Could you please point me to any
> relevant discussions on this matter?
the patchset has been posted here [1]. I didn't get around to fix the
hotremove case when you have to be extra carefule to not remove pfn
range that backs struct pages still in use. I didn't have problems for
small systems but 2GB memblocks just crashed.
[1] http://lkml.kernel.org/r/20170801124111.28881-1-mhocko@kernel.org
--
Michal Hocko
SUSE Labs
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: Memory hotplug vmem pages
2018-10-30 13:44 ` Michal Hocko
@ 2018-10-31 15:52 ` Zaslonko Mikhail
0 siblings, 0 replies; 3+ messages in thread
From: Zaslonko Mikhail @ 2018-10-31 15:52 UTC (permalink / raw)
To: Michal Hocko, Zaslonko Mikhail; +Cc: linux-mm, gerald.schaefer, schwidefsky
Hello Michal,
Thanks for your response and for the link. I'm gonna play around with that patchset.
Thanks,
Mikhail Zaslonko
On 30.10.2018 14:44, Michal Hocko wrote:
> [Sorry for late response]
>
> On Fri 12-10-18 10:15:26, Zaslonko Mikhail wrote:
>> Hello Michal,
>>
>> I've read a recent discussion about introducing the memory types for memory
>> hotplug:
>> https://marc.info/?t=153814716600004&r=1&w=2
>>
>> In particular I was interested in the idea of moving vmem struct pages to
>> the hotplugable memory itself. I'm also looking into it for s390 right now.
>> So, in one of your replies you mentioned that you "have proposed (but
>> haven't finished this due to other stuff) a solution for this". Have you
>> covered any part of that solution yet? Could you please point me to any
>> relevant discussions on this matter?
>
> the patchset has been posted here [1]. I didn't get around to fix the
> hotremove case when you have to be extra carefule to not remove pfn
> range that backs struct pages still in use. I didn't have problems for
> small systems but 2GB memblocks just crashed.
>
> [1] http://lkml.kernel.org/r/20170801124111.28881-1-mhocko@kernel.org
>
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2018-10-31 15:53 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-10-12 8:15 Memory hotplug vmem pages Zaslonko Mikhail
2018-10-30 13:44 ` Michal Hocko
2018-10-31 15:52 ` Zaslonko Mikhail
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox