Alistair Popple wrote:
> FS DAX pages have always maintained their own page reference counts
> without following the normal rules for page reference counting. In
> particular pages are considered free when the refcount hits one rather
> than zero and refcounts are not added when mapping the page.
>
Alistair Popple wrote:
> PCI P2PDMA pages are not mapped with pXX_devmap PTEs therefore the
> check in __gup_device_huge() is redundant. Remove it
>
> Signed-off-by: Alistair Popple
> Reviewed-by: Jason Gunthorpe
> Acked-by: David Hildenbrand
Acked-by: Dan Williams
> diff --git a/drivers/dax/device.c b/drivers/dax/device.c
> index eb61598..b7a31ae 100644
> --- a/drivers/dax/device.c
> +++ b/drivers/dax/device.c
> @@ -126,11 +126,11 @@ static vm_fault_t __dev_dax_pte_fault(struct dev_dax
> *dev_dax,
> return VM_FAULT_SIGBUS;
> }
>
> -
On Thu, Jun 27, 2024 at 10:54:20AM +1000, Alistair Popple wrote:
> static struct nouveau_dmem_chunk *nouveau_page_to_chunk(struct page *page)
> {
> - return container_of(page->pgmap, struct nouveau_dmem_chunk, pagemap);
> + return container_of(page_dev_pagemap(page), struct nouveau_dmem_c
On Thu, Jun 27, 2024 at 10:54:19AM +1000, Alistair Popple wrote:
> When a fs dax page is freed it has to notify filesystems that the page
> has been unpinned/unmapped and is free. Currently this involves
> special code in the page free paths to detect a transition of refcount
> from 2 to 1 and to c
On Thu, Jun 27, 2024 at 10:54:18AM +1000, Alistair Popple wrote:
> A FS DAX page is considered idle when its refcount drops to one. This
> is currently open-coded in all file systems supporting FS DAX. Move
> the idle detection to a common function to make future changes easier.
>
> Signed-off-by:
On Thu, Jun 27, 2024 at 10:54:17AM +1000, Alistair Popple wrote:
> The reference counts for ZONE_DEVICE private pages should be
> initialised by the driver when the page is actually allocated by the
> driver allocator, not when they are first created. This is currently
> the case for MEMORY_DEVICE_
On Thu, Jun 27, 2024 at 10:54:21AM +1000, Alistair Popple wrote:
> +extern void prep_compound_page(struct page *page, unsigned int order);
No need for the extern.
> static int insert_page_into_pte_locked(struct vm_area_struct *vma, pte_t
> *pte,
> - unsigned long addr, struc
On 6/26/24 8:51 PM, Thorsten Leemhuis wrote:
> On 27.06.24 01:17, Randy Dunlap wrote:
>> On 6/26/24 4:13 PM, Jonathan Corbet wrote:
>>> Konstantin Ryabitsev writes:
On Fri, Jun 21, 2024 at 02:07:44PM GMT, Kees Cook wrote:
> On Wed, Jun 19, 2024 at 02:24:07PM -0400, Konstantin Ryabitsev
On 27.06.24 01:17, Randy Dunlap wrote:
> On 6/26/24 4:13 PM, Jonathan Corbet wrote:
>> Konstantin Ryabitsev writes:
>>> On Fri, Jun 21, 2024 at 02:07:44PM GMT, Kees Cook wrote:
On Wed, Jun 19, 2024 at 02:24:07PM -0400, Konstantin Ryabitsev wrote:
> + This URL should be used when referri
Now that DAX and all other reference counts to ZONE_DEVICE pages are
managed normally there is no need for the special devmap PTE/PMD/PUD
page table bits. So drop all references to these, freeing up a
software defined page table bit on architectures supporting it.
Signed-off-by: Alistair Popple
-
The devmap PTE special bit was used to detect mappings of FS DAX
pages. This tracking was required to ensure the generic mm did not
manipulate the page reference counts as FS DAX implemented it's own
reference counting scheme.
Now that FS DAX pages have their references counted the same way as
nor
Now that DAX is managing page reference counts the same as normal
pages there are no callers for vmf_insert_pXd functions so remove
them.
Signed-off-by: Alistair Popple
---
include/linux/huge_mm.h | 2 +-
mm/huge_memory.c| 165 +-
2 files changed
Currently fs dax pages are considered free when the refcount drops to
one and their refcounts are not increased when mapped via PTEs or
decreased when unmapped. This requires special logic in mm paths to
detect that these pages should not be properly refcounted, and to
detect when the refcount drop
Longterm pinning of FS DAX pages should already be disallowed by
various pXX_devmap checks. However a future change will cause these
checks to be invalid for FS DAX pages so make
folio_is_longterm_pinnable() return false for FS DAX pages.
Signed-off-by: Alistair Popple
---
include/linux/memremap
Currently DAX folio/page reference counts are managed differently to
normal pages. To allow these to be managed the same as normal pages
introduce dax_insert_pfn_pmd. This will map the entire PMD-sized folio
and take references as it would for a normally mapped page.
This is distinct from the curr
Currently DAX folio/page reference counts are managed differently to
normal pages. To allow these to be managed the same as normal pages
introduce dax_insert_pfn_pud. This will map the entire PUD-sized folio
and take references as it would for a normally mapped page.
This is distinct from the curr
Currently to map a DAX page the DAX driver calls vmf_insert_pfn. This
creates a special devmap PTE entry for the pfn but does not take a
reference on the underlying struct page for the mapping. This is
because DAX page refcounts are treated specially, as indicated by the
presence of a devmap entry.
Zone device pages are used to represent various type of device memory
managed by device drivers. Currently compound zone device pages are
not supported. This is because MEMORY_DEVICE_FS_DAX pages are the only
user of higher order zone device pages and have their own page
reference counting.
A futu
When a fs dax page is freed it has to notify filesystems that the page
has been unpinned/unmapped and is free. Currently this involves
special code in the page free paths to detect a transition of refcount
from 2 to 1 and to call some fs dax specific code.
A future change will require this to happ
A FS DAX page is considered idle when its refcount drops to one. This
is currently open-coded in all file systems supporting FS DAX. Move
the idle detection to a common function to make future changes easier.
Signed-off-by: Alistair Popple
Reviewed-by: Jan Kara
---
fs/ext4/inode.c | 5 +
The reference counts for ZONE_DEVICE private pages should be
initialised by the driver when the page is actually allocated by the
driver allocator, not when they are first created. This is currently
the case for MEMORY_DEVICE_PRIVATE and MEMORY_DEVICE_COHERENT pages
but not MEMORY_DEVICE_PCI_P2PDMA
PCI P2PDMA pages are not mapped with pXX_devmap PTEs therefore the
check in __gup_device_huge() is redundant. Remove it
Signed-off-by: Alistair Popple
Reviewed-by: Jason Gunthorpe
Acked-by: David Hildenbrand
---
mm/gup.c | 5 -
1 file changed, 5 deletions(-)
diff --git a/mm/gup.c b/mm/gup
FS DAX pages have always maintained their own page reference counts
without following the normal rules for page reference counting. In
particular pages are considered free when the refcount hits one rather
than zero and refcounts are not added when mapping the page.
Tracking this requires special
On 6/26/24 4:13 PM, Jonathan Corbet wrote:
> Konstantin Ryabitsev writes:
>
>> On Fri, Jun 21, 2024 at 02:07:44PM GMT, Kees Cook wrote:
>>> On Wed, Jun 19, 2024 at 02:24:07PM -0400, Konstantin Ryabitsev wrote:
+ This URL should be used when referring to relevant mailing list
+ to
Konstantin Ryabitsev writes:
> On Fri, Jun 21, 2024 at 02:07:44PM GMT, Kees Cook wrote:
>> On Wed, Jun 19, 2024 at 02:24:07PM -0400, Konstantin Ryabitsev wrote:
>> > + This URL should be used when referring to relevant mailing list
>> > + topics, related patch sets, or other notable discussio
SeongJae Park writes:
> This patch series is for minor document fixups. First five patches
> clean up and remove 'Other material' section of process/index in favor
> of 'staging/' and better place for docs on the section. A patch for
> adding DAMON maintainer-profile reference on maintainers' h
On Wed, 26 Jun 2024 16:35:49 -0600 Jonathan Corbet wrote:
> SeongJae Park writes:
>
> > HacKerMaiL (hkml) [1] is a simple tool for mailing lists-based
> > development workflows such as that for most Linux kernel subsystems. It
> > is actively being maintained by DAMON maintainer, and recommend
SeongJae Park writes:
> HacKerMaiL (hkml) [1] is a simple tool for mailing lists-based
> development workflows such as that for most Linux kernel subsystems. It
> is actively being maintained by DAMON maintainer, and recommended for
> DAMON community[2]. Add a simple introduction of the tool on
SeongJae Park writes:
> 'Other material' section on 'process/index' is for unsorted documents.
> However we also have a dedicated place for the purpose, 'staging/'.
> Move 'magic-number' from the section to 'staging/' directory.
>
> Signed-off-by: SeongJae Park
> ---
> Documentation/process/ind
On Wed, Jun 26, 2024 at 10:12:35AM GMT, Geert Uytterhoeven wrote:
> > > + Link: https://patch.msgid.link/patch-source-msgid@here
> >
> > Hmm, I mentioned this in the other thread, but I also like the fact
> > that my automated script uses the list that it was Cc'd to. That is, if
> > it Cc'd li
Hi Steven,
On Tue, Jun 25, 2024 at 11:27 PM Steven Rostedt wrote:
> On Tue, 18 Jun 2024 12:42:11 -0400
> Konstantin Ryabitsev wrote:
> > + A similar approach was attempted before as part of a different
> > + effort [1], but the initial implementation caused too many
> > + regressions
32 matches
Mail list logo
