diff mbox series

[01/33] s390: Use _pt_s390_gaddr for gmap address tracking

Message ID 20230417205048.15870-2-vishal.moola@gmail.com (mailing list archive)
State New
Headers show
Series Split ptdesc from struct page | expand

Commit Message

Vishal Moola April 17, 2023, 8:50 p.m. UTC 
s390 uses page->index to keep track of page tables for the guest address
space. In an attempt to consolidate the usage of page fields in s390,
replace _pt_pad_2 with _pt_s390_gaddr to replace page->index in gmap.

This will help with the splitting of struct ptdesc from struct page, as
well as allow s390 to use _pt_frag_refcount for fragmented page table
tracking.

Since page->_pt_s390_gaddr aliases with mapping, ensure its set to NULL
before freeing the pages as well.

Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
---
 arch/s390/mm/gmap.c      | 50 +++++++++++++++++++++++++++-------------
 include/linux/mm_types.h |  2 +-
 2 files changed, 35 insertions(+), 17 deletions(-)

Comments

David Hildenbrand April 18, 2023, 3:45 p.m. UTC | #1 
On 17.04.23 22:50, Vishal Moola (Oracle) wrote:
> s390 uses page->index to keep track of page tables for the guest address
> space. In an attempt to consolidate the usage of page fields in s390,
> replace _pt_pad_2 with _pt_s390_gaddr to replace page->index in gmap.
> 
> This will help with the splitting of struct ptdesc from struct page, as
> well as allow s390 to use _pt_frag_refcount for fragmented page table
> tracking.
> 
> Since page->_pt_s390_gaddr aliases with mapping, ensure its set to NULL
> before freeing the pages as well.
> 
> Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
> ---

[...]

> diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
> index 3fc9e680f174..2616d64c0e8c 100644
> --- a/include/linux/mm_types.h
> +++ b/include/linux/mm_types.h
> @@ -144,7 +144,7 @@ struct page {
>   		struct {	/* Page table pages */
>   			unsigned long _pt_pad_1;	/* compound_head */
>   			pgtable_t pmd_huge_pte; /* protected by page->ptl */
> -			unsigned long _pt_pad_2;	/* mapping */
> +			unsigned long _pt_s390_gaddr;	/* mapping */
>   			union {
>   				struct mm_struct *pt_mm; /* x86 pgds only */
>   				atomic_t pt_frag_refcount; /* powerpc */

The confusing part is, that these gmap page tables are not ordinary 
process page tables that we would ordinarily place into this section 
here. That's why they are also not allocated/freed using the typical 
page table constructor/destructor ...
Vishal Moola April 18, 2023, 9:33 p.m. UTC | #2 
On Tue, Apr 18, 2023 at 8:45 AM David Hildenbrand <david@redhat.com> wrote:
>
> On 17.04.23 22:50, Vishal Moola (Oracle) wrote:
> > s390 uses page->index to keep track of page tables for the guest address
> > space. In an attempt to consolidate the usage of page fields in s390,
> > replace _pt_pad_2 with _pt_s390_gaddr to replace page->index in gmap.
> >
> > This will help with the splitting of struct ptdesc from struct page, as
> > well as allow s390 to use _pt_frag_refcount for fragmented page table
> > tracking.
> >
> > Since page->_pt_s390_gaddr aliases with mapping, ensure its set to NULL
> > before freeing the pages as well.
> >
> > Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
> > ---
>
> [...]
>
> > diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
> > index 3fc9e680f174..2616d64c0e8c 100644
> > --- a/include/linux/mm_types.h
> > +++ b/include/linux/mm_types.h
> > @@ -144,7 +144,7 @@ struct page {
> >               struct {        /* Page table pages */
> >                       unsigned long _pt_pad_1;        /* compound_head */
> >                       pgtable_t pmd_huge_pte; /* protected by page->ptl */
> > -                     unsigned long _pt_pad_2;        /* mapping */
> > +                     unsigned long _pt_s390_gaddr;   /* mapping */
> >                       union {
> >                               struct mm_struct *pt_mm; /* x86 pgds only */
> >                               atomic_t pt_frag_refcount; /* powerpc */
>
> The confusing part is, that these gmap page tables are not ordinary
> process page tables that we would ordinarily place into this section
> here. That's why they are also not allocated/freed using the typical
> page table constructor/destructor ...

I initially thought the same, so I was quite confused when I saw
__gmap_segment_gaddr was using pmd_pgtable_page().

Although they are not ordinary process page tables, since we
eventually want to move them out of struct page, I think shifting them
to be in ptdescs, being a memory descriptor for page tables, makes
the most sense.

Another option is to leave pmd_pgtable_page() as is just for this case.
Or we can revert commit 7e25de77bc5ea which uses the function here
then figure out where these gmap pages table pages will go later.
David Hildenbrand April 19, 2023, 7:54 a.m. UTC | #3 
On 18.04.23 23:33, Vishal Moola wrote:
> On Tue, Apr 18, 2023 at 8:45 AM David Hildenbrand <david@redhat.com> wrote:
>>
>> On 17.04.23 22:50, Vishal Moola (Oracle) wrote:
>>> s390 uses page->index to keep track of page tables for the guest address
>>> space. In an attempt to consolidate the usage of page fields in s390,
>>> replace _pt_pad_2 with _pt_s390_gaddr to replace page->index in gmap.
>>>
>>> This will help with the splitting of struct ptdesc from struct page, as
>>> well as allow s390 to use _pt_frag_refcount for fragmented page table
>>> tracking.
>>>
>>> Since page->_pt_s390_gaddr aliases with mapping, ensure its set to NULL
>>> before freeing the pages as well.
>>>
>>> Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
>>> ---
>>
>> [...]
>>
>>> diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
>>> index 3fc9e680f174..2616d64c0e8c 100644
>>> --- a/include/linux/mm_types.h
>>> +++ b/include/linux/mm_types.h
>>> @@ -144,7 +144,7 @@ struct page {
>>>                struct {        /* Page table pages */
>>>                        unsigned long _pt_pad_1;        /* compound_head */
>>>                        pgtable_t pmd_huge_pte; /* protected by page->ptl */
>>> -                     unsigned long _pt_pad_2;        /* mapping */
>>> +                     unsigned long _pt_s390_gaddr;   /* mapping */
>>>                        union {
>>>                                struct mm_struct *pt_mm; /* x86 pgds only */
>>>                                atomic_t pt_frag_refcount; /* powerpc */
>>
>> The confusing part is, that these gmap page tables are not ordinary
>> process page tables that we would ordinarily place into this section
>> here. That's why they are also not allocated/freed using the typical
>> page table constructor/destructor ...
> 
> I initially thought the same, so I was quite confused when I saw
> __gmap_segment_gaddr was using pmd_pgtable_page().
> 
> Although they are not ordinary process page tables, since we
> eventually want to move them out of struct page, I think shifting them
> to be in ptdescs, being a memory descriptor for page tables, makes
> the most sense.

Seeing utilities like tlb_remove_page_ptdesc() that don't really apply 
to such page tables, I wonder if we should much rather treat such 
shadow/auxiliary/... page tables (just like other architectures like 
x86, arm, ... employ as well) as a distinct type.

And have ptdesc be the common type for all process page tables.

> 
> Another option is to leave pmd_pgtable_page() as is just for this case.
> Or we can revert commit 7e25de77bc5ea which uses the function here
> then figure out where these gmap pages table pages will go later.

I'm always confused when reading gmap code, so let me have another look :)

The confusing part is that s390x shares the lowest level page tables 
(PTE tables) between the process and gmap ("guest mapping", similar to 
EPT on x86-64). It maps these process PTE tables (covering 1 MiB) into 
gmap-specific PMD tables.

pmd_pgtable_page() should indeed always give us a gmap-specific 
PMD-table. In fact, something allocated via gmap_alloc_table().

Decoupling both concepts sounds like a good idea.
Vishal Moola April 20, 2023, 11:32 p.m. UTC | #4 
On Wed, Apr 19, 2023 at 12:54 AM David Hildenbrand <david@redhat.com> wrote:
>
> On 18.04.23 23:33, Vishal Moola wrote:
> > On Tue, Apr 18, 2023 at 8:45 AM David Hildenbrand <david@redhat.com> wrote:
> >>
> >> On 17.04.23 22:50, Vishal Moola (Oracle) wrote:
> >>> s390 uses page->index to keep track of page tables for the guest address
> >>> space. In an attempt to consolidate the usage of page fields in s390,
> >>> replace _pt_pad_2 with _pt_s390_gaddr to replace page->index in gmap.
> >>>
> >>> This will help with the splitting of struct ptdesc from struct page, as
> >>> well as allow s390 to use _pt_frag_refcount for fragmented page table
> >>> tracking.
> >>>
> >>> Since page->_pt_s390_gaddr aliases with mapping, ensure its set to NULL
> >>> before freeing the pages as well.
> >>>
> >>> Signed-off-by: Vishal Moola (Oracle) <vishal.moola@gmail.com>
> >>> ---
> >>
> >> [...]
> >>
> >>> diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
> >>> index 3fc9e680f174..2616d64c0e8c 100644
> >>> --- a/include/linux/mm_types.h
> >>> +++ b/include/linux/mm_types.h
> >>> @@ -144,7 +144,7 @@ struct page {
> >>>                struct {        /* Page table pages */
> >>>                        unsigned long _pt_pad_1;        /* compound_head */
> >>>                        pgtable_t pmd_huge_pte; /* protected by page->ptl */
> >>> -                     unsigned long _pt_pad_2;        /* mapping */
> >>> +                     unsigned long _pt_s390_gaddr;   /* mapping */
> >>>                        union {
> >>>                                struct mm_struct *pt_mm; /* x86 pgds only */
> >>>                                atomic_t pt_frag_refcount; /* powerpc */
> >>
> >> The confusing part is, that these gmap page tables are not ordinary
> >> process page tables that we would ordinarily place into this section
> >> here. That's why they are also not allocated/freed using the typical
> >> page table constructor/destructor ...
> >
> > I initially thought the same, so I was quite confused when I saw
> > __gmap_segment_gaddr was using pmd_pgtable_page().
> >
> > Although they are not ordinary process page tables, since we
> > eventually want to move them out of struct page, I think shifting them
> > to be in ptdescs, being a memory descriptor for page tables, makes
> > the most sense.
>
> Seeing utilities like tlb_remove_page_ptdesc() that don't really apply
> to such page tables, I wonder if we should much rather treat such
> shadow/auxiliary/... page tables (just like other architectures like
> x86, arm, ... employ as well) as a distinct type.
>
> And have ptdesc be the common type for all process page tables.

Although I do like the idea of having a distinct type for them, I'm not sure
I see the merits of having another type specifically for those types of
page tables.

As it currently is, tlb_remove_table() is only distinct from tlb_remove_page()
when an architecture defines its own removal function. I'm not too familiar
with most of their differences, but we can probably continue to let them do
that. As of now, I'm not too sure what a distinct type would look like that
could meet all their needs holistically.

> >
> > Another option is to leave pmd_pgtable_page() as is just for this case.
> > Or we can revert commit 7e25de77bc5ea which uses the function here
> > then figure out where these gmap pages table pages will go later.
>
> I'm always confused when reading gmap code, so let me have another look :)
>
> The confusing part is that s390x shares the lowest level page tables
> (PTE tables) between the process and gmap ("guest mapping", similar to
> EPT on x86-64). It maps these process PTE tables (covering 1 MiB) into
> gmap-specific PMD tables.

Especially in cases like this. If the architecture wants to share page tables
then everything being in form ptdesc would make that easiest, and
continue to let them define their own niche functions for their needs.

> pmd_pgtable_page() should indeed always give us a gmap-specific
> PMD-table. In fact, something allocated via gmap_alloc_table().
>
> Decoupling both concepts sounds like a good idea.

Yeah, I'm not a fan of how this gmap caller is the only external caller
using this to get a page for their own purposes. I'll update that in v2.
diff mbox series

Patch

diff --git a/arch/s390/mm/gmap.c b/arch/s390/mm/gmap.c
index 5a716bdcba05..a61ea1a491dc 100644
--- a/arch/s390/mm/gmap.c
+++ b/arch/s390/mm/gmap.c
@@ -70,7 +70,7 @@  static struct gmap *gmap_alloc(unsigned long limit)
 	page = alloc_pages(GFP_KERNEL_ACCOUNT, CRST_ALLOC_ORDER);
 	if (!page)
 		goto out_free;
-	page->index = 0;
+	page->_pt_s390_gaddr = 0;
 	list_add(&page->lru, &gmap->crst_list);
 	table = page_to_virt(page);
 	crst_table_init(table, etype);
@@ -187,16 +187,20 @@  static void gmap_free(struct gmap *gmap)
 	if (!(gmap_is_shadow(gmap) && gmap->removed))
 		gmap_flush_tlb(gmap);
 	/* Free all segment & region tables. */
-	list_for_each_entry_safe(page, next, &gmap->crst_list, lru)
+	list_for_each_entry_safe(page, next, &gmap->crst_list, lru) {
+		page->_pt_s390_gaddr = 0;
 		__free_pages(page, CRST_ALLOC_ORDER);
+	}
 	gmap_radix_tree_free(&gmap->guest_to_host);
 	gmap_radix_tree_free(&gmap->host_to_guest);
 
 	/* Free additional data for a shadow gmap */
 	if (gmap_is_shadow(gmap)) {
 		/* Free all page tables. */
-		list_for_each_entry_safe(page, next, &gmap->pt_list, lru)
+		list_for_each_entry_safe(page, next, &gmap->pt_list, lru) {
+			page->_pt_s390_gaddr = 0;
 			page_table_free_pgste(page);
+		}
 		gmap_rmap_radix_tree_free(&gmap->host_to_rmap);
 		/* Release reference to the parent */
 		gmap_put(gmap->parent);
@@ -318,12 +322,14 @@  static int gmap_alloc_table(struct gmap *gmap, unsigned long *table,
 		list_add(&page->lru, &gmap->crst_list);
 		*table = __pa(new) | _REGION_ENTRY_LENGTH |
 			(*table & _REGION_ENTRY_TYPE_MASK);
-		page->index = gaddr;
+		page->_pt_s390_gaddr = gaddr;
 		page = NULL;
 	}
 	spin_unlock(&gmap->guest_table_lock);
-	if (page)
+	if (page) {
+		page->_pt_s390_gaddr = 0;
 		__free_pages(page, CRST_ALLOC_ORDER);
+	}
 	return 0;
 }
 
@@ -341,7 +347,7 @@  static unsigned long __gmap_segment_gaddr(unsigned long *entry)
 	offset = (unsigned long) entry / sizeof(unsigned long);
 	offset = (offset & (PTRS_PER_PMD - 1)) * PMD_SIZE;
 	page = pmd_pgtable_page((pmd_t *) entry);
-	return page->index + offset;
+	return page->_pt_s390_gaddr + offset;
 }
 
 /**
@@ -1351,6 +1357,7 @@  static void gmap_unshadow_pgt(struct gmap *sg, unsigned long raddr)
 	/* Free page table */
 	page = phys_to_page(pgt);
 	list_del(&page->lru);
+	page->_pt_s390_gaddr = 0;
 	page_table_free_pgste(page);
 }
 
@@ -1379,6 +1386,7 @@  static void __gmap_unshadow_sgt(struct gmap *sg, unsigned long raddr,
 		/* Free page table */
 		page = phys_to_page(pgt);
 		list_del(&page->lru);
+		page->_pt_s390_gaddr = 0;
 		page_table_free_pgste(page);
 	}
 }
@@ -1409,6 +1417,7 @@  static void gmap_unshadow_sgt(struct gmap *sg, unsigned long raddr)
 	/* Free segment table */
 	page = phys_to_page(sgt);
 	list_del(&page->lru);
+	page->_pt_s390_gaddr = 0;
 	__free_pages(page, CRST_ALLOC_ORDER);
 }
 
@@ -1437,6 +1446,7 @@  static void __gmap_unshadow_r3t(struct gmap *sg, unsigned long raddr,
 		/* Free segment table */
 		page = phys_to_page(sgt);
 		list_del(&page->lru);
+		page->_pt_s390_gaddr = 0;
 		__free_pages(page, CRST_ALLOC_ORDER);
 	}
 }
@@ -1467,6 +1477,7 @@  static void gmap_unshadow_r3t(struct gmap *sg, unsigned long raddr)
 	/* Free region 3 table */
 	page = phys_to_page(r3t);
 	list_del(&page->lru);
+	page->_pt_s390_gaddr = 0;
 	__free_pages(page, CRST_ALLOC_ORDER);
 }
 
@@ -1495,6 +1506,7 @@  static void __gmap_unshadow_r2t(struct gmap *sg, unsigned long raddr,
 		/* Free region 3 table */
 		page = phys_to_page(r3t);
 		list_del(&page->lru);
+		page->_pt_s390_gaddr = 0;
 		__free_pages(page, CRST_ALLOC_ORDER);
 	}
 }
@@ -1525,6 +1537,7 @@  static void gmap_unshadow_r2t(struct gmap *sg, unsigned long raddr)
 	/* Free region 2 table */
 	page = phys_to_page(r2t);
 	list_del(&page->lru);
+	page->_pt_s390_gaddr = 0;
 	__free_pages(page, CRST_ALLOC_ORDER);
 }
 
@@ -1557,6 +1570,7 @@  static void __gmap_unshadow_r1t(struct gmap *sg, unsigned long raddr,
 		/* Free region 2 table */
 		page = phys_to_page(r2t);
 		list_del(&page->lru);
+		page->_pt_s390_gaddr = 0;
 		__free_pages(page, CRST_ALLOC_ORDER);
 	}
 }
@@ -1762,9 +1776,9 @@  int gmap_shadow_r2t(struct gmap *sg, unsigned long saddr, unsigned long r2t,
 	page = alloc_pages(GFP_KERNEL_ACCOUNT, CRST_ALLOC_ORDER);
 	if (!page)
 		return -ENOMEM;
-	page->index = r2t & _REGION_ENTRY_ORIGIN;
+	page->_pt_s390_gaddr = r2t & _REGION_ENTRY_ORIGIN;
 	if (fake)
-		page->index |= GMAP_SHADOW_FAKE_TABLE;
+		page->_pt_s390_gaddr |= GMAP_SHADOW_FAKE_TABLE;
 	s_r2t = page_to_phys(page);
 	/* Install shadow region second table */
 	spin_lock(&sg->guest_table_lock);
@@ -1814,6 +1828,7 @@  int gmap_shadow_r2t(struct gmap *sg, unsigned long saddr, unsigned long r2t,
 	return rc;
 out_free:
 	spin_unlock(&sg->guest_table_lock);
+	page->_pt_s390_gaddr = 0;
 	__free_pages(page, CRST_ALLOC_ORDER);
 	return rc;
 }
@@ -1846,9 +1861,9 @@  int gmap_shadow_r3t(struct gmap *sg, unsigned long saddr, unsigned long r3t,
 	page = alloc_pages(GFP_KERNEL_ACCOUNT, CRST_ALLOC_ORDER);
 	if (!page)
 		return -ENOMEM;
-	page->index = r3t & _REGION_ENTRY_ORIGIN;
+	page->_pt_s390_gaddr = r3t & _REGION_ENTRY_ORIGIN;
 	if (fake)
-		page->index |= GMAP_SHADOW_FAKE_TABLE;
+		page->_pt_s390_gaddr |= GMAP_SHADOW_FAKE_TABLE;
 	s_r3t = page_to_phys(page);
 	/* Install shadow region second table */
 	spin_lock(&sg->guest_table_lock);
@@ -1898,6 +1913,7 @@  int gmap_shadow_r3t(struct gmap *sg, unsigned long saddr, unsigned long r3t,
 	return rc;
 out_free:
 	spin_unlock(&sg->guest_table_lock);
+	page->_pt_s390_gaddr = 0;
 	__free_pages(page, CRST_ALLOC_ORDER);
 	return rc;
 }
@@ -1930,9 +1946,9 @@  int gmap_shadow_sgt(struct gmap *sg, unsigned long saddr, unsigned long sgt,
 	page = alloc_pages(GFP_KERNEL_ACCOUNT, CRST_ALLOC_ORDER);
 	if (!page)
 		return -ENOMEM;
-	page->index = sgt & _REGION_ENTRY_ORIGIN;
+	page->_pt_s390_gaddr = sgt & _REGION_ENTRY_ORIGIN;
 	if (fake)
-		page->index |= GMAP_SHADOW_FAKE_TABLE;
+		page->_pt_s390_gaddr |= GMAP_SHADOW_FAKE_TABLE;
 	s_sgt = page_to_phys(page);
 	/* Install shadow region second table */
 	spin_lock(&sg->guest_table_lock);
@@ -1982,6 +1998,7 @@  int gmap_shadow_sgt(struct gmap *sg, unsigned long saddr, unsigned long sgt,
 	return rc;
 out_free:
 	spin_unlock(&sg->guest_table_lock);
+	page->_pt_s390_gaddr = 0;
 	__free_pages(page, CRST_ALLOC_ORDER);
 	return rc;
 }
@@ -2014,9 +2031,9 @@  int gmap_shadow_pgt_lookup(struct gmap *sg, unsigned long saddr,
 	if (table && !(*table & _SEGMENT_ENTRY_INVALID)) {
 		/* Shadow page tables are full pages (pte+pgste) */
 		page = pfn_to_page(*table >> PAGE_SHIFT);
-		*pgt = page->index & ~GMAP_SHADOW_FAKE_TABLE;
+		*pgt = page->_pt_s390_gaddr & ~GMAP_SHADOW_FAKE_TABLE;
 		*dat_protection = !!(*table & _SEGMENT_ENTRY_PROTECT);
-		*fake = !!(page->index & GMAP_SHADOW_FAKE_TABLE);
+		*fake = !!(page->_pt_s390_gaddr & GMAP_SHADOW_FAKE_TABLE);
 		rc = 0;
 	} else  {
 		rc = -EAGAIN;
@@ -2054,9 +2071,9 @@  int gmap_shadow_pgt(struct gmap *sg, unsigned long saddr, unsigned long pgt,
 	page = page_table_alloc_pgste(sg->mm);
 	if (!page)
 		return -ENOMEM;
-	page->index = pgt & _SEGMENT_ENTRY_ORIGIN;
+	page->_pt_s390_gaddr = pgt & _SEGMENT_ENTRY_ORIGIN;
 	if (fake)
-		page->index |= GMAP_SHADOW_FAKE_TABLE;
+		page->_pt_s390_gaddr |= GMAP_SHADOW_FAKE_TABLE;
 	s_pgt = page_to_phys(page);
 	/* Install shadow page table */
 	spin_lock(&sg->guest_table_lock);
@@ -2101,6 +2118,7 @@  int gmap_shadow_pgt(struct gmap *sg, unsigned long saddr, unsigned long pgt,
 	return rc;
 out_free:
 	spin_unlock(&sg->guest_table_lock);
+	page->_pt_s390_gaddr = 0;
 	page_table_free_pgste(page);
 	return rc;
 
diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h
index 3fc9e680f174..2616d64c0e8c 100644
--- a/include/linux/mm_types.h
+++ b/include/linux/mm_types.h
@@ -144,7 +144,7 @@  struct page {
 		struct {	/* Page table pages */
 			unsigned long _pt_pad_1;	/* compound_head */
 			pgtable_t pmd_huge_pte; /* protected by page->ptl */
-			unsigned long _pt_pad_2;	/* mapping */
+			unsigned long _pt_s390_gaddr;	/* mapping */
 			union {
 				struct mm_struct *pt_mm; /* x86 pgds only */
 				atomic_t pt_frag_refcount; /* powerpc */