Re: [PATCH v6 08/14] mm/gup: do not migrate zero page

2021-01-25 Thread Jason Gunthorpe
On Wed, Jan 20, 2021 at 09:26:41AM -0500, Pavel Tatashin wrote: > I thought about this, and it would code a little cleaner. But, the > reason I did not is because zero_page is perfectly pinnable, it is not > pinnable only when it is in a movable zone (and it should not be in a > movable zones for

Re: [PATCH v6 08/14] mm/gup: do not migrate zero page

2021-01-25 Thread Pavel Tatashin
On Mon, Jan 25, 2021 at 9:28 AM Jason Gunthorpe wrote: > > On Wed, Jan 20, 2021 at 09:26:41AM -0500, Pavel Tatashin wrote: > > > I thought about this, and it would code a little cleaner. But, the > > reason I did not is because zero_page is perfectly pinnable, it is not > > pinnable only when it

Re: [PATCH v6 08/14] mm/gup: do not migrate zero page

2021-01-20 Thread Pavel Tatashin
On Wed, Jan 20, 2021 at 8:14 AM Jason Gunthorpe wrote: > > On Tue, Jan 19, 2021 at 08:43:27PM -0500, Pavel Tatashin wrote: > > On some platforms ZERO_PAGE(0) might end-up in a movable zone. Do not > > migrate zero page in gup during longterm pinning as migration of zero page > > is not allowed. >

Re: [PATCH v6 08/14] mm/gup: do not migrate zero page

2021-01-20 Thread Jason Gunthorpe
On Tue, Jan 19, 2021 at 08:43:27PM -0500, Pavel Tatashin wrote: > On some platforms ZERO_PAGE(0) might end-up in a movable zone. Do not > migrate zero page in gup during longterm pinning as migration of zero page > is not allowed. > > For example, in x86 QEMU with 16G of memory and kernelcore=5G

[PATCH v6 08/14] mm/gup: do not migrate zero page

2021-01-19 Thread Pavel Tatashin
On some platforms ZERO_PAGE(0) might end-up in a movable zone. Do not migrate zero page in gup during longterm pinning as migration of zero page is not allowed. For example, in x86 QEMU with 16G of memory and kernelcore=5G parameter, I see the following: Boot#1: zero_pfn 0x48a8d zero_pfn zone: