Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-05 Thread David Hildenbrand
On 04.06.20 22:00, Dave Hansen wrote: > On 6/4/20 11:12 AM, Daniel Jordan wrote: >>> E.g., on powerpc that's 16MB so they have *a lot* of memory blocks. >>> That's why that's not papering over the problem. Increasing the memory >>> block size isn't always the answer. >> Ok. If you don't mind, what

Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-04 Thread Daniel Jordan
On Thu, Jun 04, 2020 at 01:00:55PM -0700, Dave Hansen wrote: > On 6/4/20 11:12 AM, Daniel Jordan wrote: > >> E.g., on powerpc that's 16MB so they have *a lot* of memory blocks. > >> That's why that's not papering over the problem. Increasing the memory > >> block size isn't always the answer. > > O

Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-04 Thread Daniel Jordan
On Thu, Jun 04, 2020 at 08:55:19PM +0200, David Hildenbrand wrote: > >> E.g., on powerpc that's 16MB so they have *a lot* of memory blocks. > >> That's why that's not papering over the problem. Increasing the memory > >> block size isn't always the answer. > > > > Ok. If you don't mind, what's th

Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-04 Thread Dave Hansen
On 6/4/20 11:12 AM, Daniel Jordan wrote: >> E.g., on powerpc that's 16MB so they have *a lot* of memory blocks. >> That's why that's not papering over the problem. Increasing the memory >> block size isn't always the answer. > Ok. If you don't mind, what's the purpose of hotplugging at that granul

Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-04 Thread David Hildenbrand
>> E.g., on powerpc that's 16MB so they have *a lot* of memory blocks. >> That's why that's not papering over the problem. Increasing the memory >> block size isn't always the answer. > > Ok. If you don't mind, what's the purpose of hotplugging at that granularity? > I'm simply curious. On bare

Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-04 Thread Daniel Jordan
On Thu, Jun 04, 2020 at 07:45:40PM +0200, David Hildenbrand wrote: > On 04.06.20 19:22, Daniel Jordan wrote: > > IMHO the root cause of this is really the small block size. Building a > > cache > > on top to avoid iterating over tons of small blocks seems like papering over > > the problem, espec

Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-04 Thread David Hildenbrand
On 04.06.20 19:22, Daniel Jordan wrote: > On Thu, Jun 04, 2020 at 09:22:03AM +0200, David Hildenbrand wrote: >> On 04.06.20 05:54, Daniel Jordan wrote: >>> Some of our servers spend 14 out of the 21 seconds of kernel boot >>> initializing memory block sysfs directories and then creating symlinks >>

Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-04 Thread Daniel Jordan
On Thu, Jun 04, 2020 at 09:22:03AM +0200, David Hildenbrand wrote: > On 04.06.20 05:54, Daniel Jordan wrote: > > Some of our servers spend 14 out of the 21 seconds of kernel boot > > initializing memory block sysfs directories and then creating symlinks > > between them and the corresponding nodes.

Re: [PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-04 Thread David Hildenbrand
On 04.06.20 05:54, Daniel Jordan wrote: > Some of our servers spend 14 out of the 21 seconds of kernel boot > initializing memory block sysfs directories and then creating symlinks > between them and the corresponding nodes. The slowness happens because > the machines get stuck with the smallest s

[PATCH] x86/mm: use max memory block size with unaligned memory end

2020-06-03 Thread Daniel Jordan
Some of our servers spend 14 out of the 21 seconds of kernel boot initializing memory block sysfs directories and then creating symlinks between them and the corresponding nodes. The slowness happens because the machines get stuck with the smallest supported memory block size on x86 (128M), which