* [PATCH v2 0/2] swap readahead clean up
@ 2017-11-01 7:15 ` Minchan Kim
0 siblings, 0 replies; 6+ messages in thread
From: Minchan Kim @ 2017-11-01 7:15 UTC (permalink / raw)
To: Andrew Morton
Cc: linux-kernel, linux-mm, Hugh Dickins, Huang Ying, kernel-team,
Minchan Kim
This patchset cleans up recent added vma-based readahead code via
unifying cluster-based readahead.
* From v1
* renamed swap_cluster_readahead - Huang
* remove unnecessary extern declaration - Huang
* add description for swapin_readahead - Huang
Minchan Kim (2):
mm:swap: clean up swap readahead
mm:swap: unify cluster-based and vma-based swap readahead
include/linux/swap.h | 38 +++-----------
mm/memory.c | 24 +++------
mm/shmem.c | 5 +-
mm/swap_state.c | 137 ++++++++++++++++++++++++++++++++-------------------
4 files changed, 103 insertions(+), 101 deletions(-)
--
2.7.4
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v2 1/2] mm:swap: clean up swap readahead
2017-11-01 7:15 ` Minchan Kim
@ 2017-11-01 7:15 ` Minchan Kim
-1 siblings, 0 replies; 6+ messages in thread
From: Minchan Kim @ 2017-11-01 7:15 UTC (permalink / raw)
To: Andrew Morton
Cc: linux-kernel, linux-mm, Hugh Dickins, Huang Ying, kernel-team,
Minchan Kim
When I see recent change of swap readahead, I am very unhappy
about current code structure which diverges two swap readahead
algorithm in do_swap_page. This patch is to clean it up.
Main motivation is that fault handler doesn't need to be aware of
readahead algorithms but just should call swapin_readahead.
As first step, this patch cleans up a little bit but not perfect
(I just separate for review easier) so next patch will make the goal
complete.
Signed-off-by: Minchan Kim <minchan@kernel.org>
---
include/linux/swap.h | 17 ++--------
mm/memory.c | 17 +++-------
mm/swap_state.c | 89 ++++++++++++++++++++++++++++------------------------
3 files changed, 55 insertions(+), 68 deletions(-)
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 84255b3da7c1..7c7c8b344bc9 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -427,12 +427,8 @@ extern struct page *__read_swap_cache_async(swp_entry_t, gfp_t,
bool *new_page_allocated);
extern struct page *swapin_readahead(swp_entry_t, gfp_t,
struct vm_area_struct *vma, unsigned long addr);
-
-extern struct page *swap_readahead_detect(struct vm_fault *vmf,
- struct vma_swap_readahead *swap_ra);
extern struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
- struct vm_fault *vmf,
- struct vma_swap_readahead *swap_ra);
+ struct vm_fault *vmf);
/* linux/mm/swapfile.c */
extern atomic_long_t nr_swap_pages;
@@ -551,15 +547,8 @@ static inline bool swap_use_vma_readahead(void)
return false;
}
-static inline struct page *swap_readahead_detect(
- struct vm_fault *vmf, struct vma_swap_readahead *swap_ra)
-{
- return NULL;
-}
-
-static inline struct page *do_swap_page_readahead(
- swp_entry_t fentry, gfp_t gfp_mask,
- struct vm_fault *vmf, struct vma_swap_readahead *swap_ra)
+static inline struct page *do_swap_page_readahead(swp_entry_t fentry,
+ gfp_t gfp_mask, struct vm_fault *vmf)
{
return NULL;
}
diff --git a/mm/memory.c b/mm/memory.c
index 8a0c410037d2..e955298e4290 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2849,21 +2849,14 @@ int do_swap_page(struct vm_fault *vmf)
struct vm_area_struct *vma = vmf->vma;
struct page *page = NULL, *swapcache = NULL;
struct mem_cgroup *memcg;
- struct vma_swap_readahead swap_ra;
swp_entry_t entry;
pte_t pte;
int locked;
int exclusive = 0;
int ret = 0;
- bool vma_readahead = swap_use_vma_readahead();
- if (vma_readahead)
- page = swap_readahead_detect(vmf, &swap_ra);
- if (!pte_unmap_same(vma->vm_mm, vmf->pmd, vmf->pte, vmf->orig_pte)) {
- if (page)
- put_page(page);
+ if (!pte_unmap_same(vma->vm_mm, vmf->pmd, vmf->pte, vmf->orig_pte))
goto out;
- }
entry = pte_to_swp_entry(vmf->orig_pte);
if (unlikely(non_swap_entry(entry))) {
@@ -2889,9 +2882,7 @@ int do_swap_page(struct vm_fault *vmf)
delayacct_set_flag(DELAYACCT_PF_SWAPIN);
- if (!page)
- page = lookup_swap_cache(entry, vma_readahead ? vma : NULL,
- vmf->address);
+ page = lookup_swap_cache(entry, vma, vmf->address);
if (!page) {
struct swap_info_struct *si = swp_swap_info(entry);
@@ -2907,9 +2898,9 @@ int do_swap_page(struct vm_fault *vmf)
swap_readpage(page, true);
}
} else {
- if (vma_readahead)
+ if (swap_use_vma_readahead())
page = do_swap_page_readahead(entry,
- GFP_HIGHUSER_MOVABLE, vmf, &swap_ra);
+ GFP_HIGHUSER_MOVABLE, vmf);
else
page = swapin_readahead(entry,
GFP_HIGHUSER_MOVABLE, vma, vmf->address);
diff --git a/mm/swap_state.c b/mm/swap_state.c
index 6c017ced11e6..e3c535fcd2df 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -331,32 +331,38 @@ struct page *lookup_swap_cache(swp_entry_t entry, struct vm_area_struct *vma,
unsigned long addr)
{
struct page *page;
- unsigned long ra_info;
- int win, hits, readahead;
page = find_get_page(swap_address_space(entry), swp_offset(entry));
INC_CACHE_INFO(find_total);
if (page) {
+ bool vma_ra = swap_use_vma_readahead();
+ bool readahead = TestClearPageReadahead(page);
+
INC_CACHE_INFO(find_success);
if (unlikely(PageTransCompound(page)))
return page;
- readahead = TestClearPageReadahead(page);
- if (vma) {
- ra_info = GET_SWAP_RA_VAL(vma);
- win = SWAP_RA_WIN(ra_info);
- hits = SWAP_RA_HITS(ra_info);
+
+ if (vma && vma_ra) {
+ unsigned long ra_val;
+ int win, hits;
+
+ ra_val = GET_SWAP_RA_VAL(vma);
+ win = SWAP_RA_WIN(ra_val);
+ hits = SWAP_RA_HITS(ra_val);
if (readahead)
hits = min_t(int, hits + 1, SWAP_RA_HITS_MAX);
atomic_long_set(&vma->swap_readahead_info,
SWAP_RA_VAL(addr, win, hits));
}
+
if (readahead) {
count_vm_event(SWAP_RA_HIT);
- if (!vma)
+ if (!vma || !vma_ra)
atomic_inc(&swapin_readahead_hits);
}
}
+
return page;
}
@@ -645,16 +651,15 @@ static inline void swap_ra_clamp_pfn(struct vm_area_struct *vma,
PFN_DOWN((faddr & PMD_MASK) + PMD_SIZE));
}
-struct page *swap_readahead_detect(struct vm_fault *vmf,
- struct vma_swap_readahead *swap_ra)
+static void swap_ra_info(struct vm_fault *vmf,
+ struct vma_swap_readahead *ra_info)
{
struct vm_area_struct *vma = vmf->vma;
- unsigned long swap_ra_info;
- struct page *page;
+ unsigned long ra_val;
swp_entry_t entry;
unsigned long faddr, pfn, fpfn;
unsigned long start, end;
- pte_t *pte;
+ pte_t *pte, *orig_pte;
unsigned int max_win, hits, prev_win, win, left;
#ifndef CONFIG_64BIT
pte_t *tpte;
@@ -663,30 +668,32 @@ struct page *swap_readahead_detect(struct vm_fault *vmf,
max_win = 1 << min_t(unsigned int, READ_ONCE(page_cluster),
SWAP_RA_ORDER_CEILING);
if (max_win == 1) {
- swap_ra->win = 1;
- return NULL;
+ ra_info->win = 1;
+ return;
}
faddr = vmf->address;
- entry = pte_to_swp_entry(vmf->orig_pte);
- if ((unlikely(non_swap_entry(entry))))
- return NULL;
- page = lookup_swap_cache(entry, vma, faddr);
- if (page)
- return page;
+ orig_pte = pte = pte_offset_map(vmf->pmd, faddr);
+ entry = pte_to_swp_entry(*pte);
+ if ((unlikely(non_swap_entry(entry)))) {
+ pte_unmap(orig_pte);
+ return;
+ }
fpfn = PFN_DOWN(faddr);
- swap_ra_info = GET_SWAP_RA_VAL(vma);
- pfn = PFN_DOWN(SWAP_RA_ADDR(swap_ra_info));
- prev_win = SWAP_RA_WIN(swap_ra_info);
- hits = SWAP_RA_HITS(swap_ra_info);
- swap_ra->win = win = __swapin_nr_pages(pfn, fpfn, hits,
+ ra_val = GET_SWAP_RA_VAL(vma);
+ pfn = PFN_DOWN(SWAP_RA_ADDR(ra_val));
+ prev_win = SWAP_RA_WIN(ra_val);
+ hits = SWAP_RA_HITS(ra_val);
+ ra_info->win = win = __swapin_nr_pages(pfn, fpfn, hits,
max_win, prev_win);
atomic_long_set(&vma->swap_readahead_info,
SWAP_RA_VAL(faddr, win, 0));
- if (win == 1)
- return NULL;
+ if (win == 1) {
+ pte_unmap(orig_pte);
+ return;
+ }
/* Copy the PTEs because the page table may be unmapped */
if (fpfn == pfn + 1)
@@ -699,23 +706,21 @@ struct page *swap_readahead_detect(struct vm_fault *vmf,
swap_ra_clamp_pfn(vma, faddr, fpfn - left, fpfn + win - left,
&start, &end);
}
- swap_ra->nr_pte = end - start;
- swap_ra->offset = fpfn - start;
- pte = vmf->pte - swap_ra->offset;
+ ra_info->nr_pte = end - start;
+ ra_info->offset = fpfn - start;
+ pte -= ra_info->offset;
#ifdef CONFIG_64BIT
- swap_ra->ptes = pte;
+ ra_info->ptes = pte;
#else
- tpte = swap_ra->ptes;
+ tpte = ra_info->ptes;
for (pfn = start; pfn != end; pfn++)
*tpte++ = *pte++;
#endif
-
- return NULL;
+ pte_unmap(orig_pte);
}
struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
- struct vm_fault *vmf,
- struct vma_swap_readahead *swap_ra)
+ struct vm_fault *vmf)
{
struct blk_plug plug;
struct vm_area_struct *vma = vmf->vma;
@@ -724,12 +729,14 @@ struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
swp_entry_t entry;
unsigned int i;
bool page_allocated;
+ struct vma_swap_readahead ra_info = {0,};
- if (swap_ra->win == 1)
+ swap_ra_info(vmf, &ra_info);
+ if (ra_info.win == 1)
goto skip;
blk_start_plug(&plug);
- for (i = 0, pte = swap_ra->ptes; i < swap_ra->nr_pte;
+ for (i = 0, pte = ra_info.ptes; i < ra_info.nr_pte;
i++, pte++) {
pentry = *pte;
if (pte_none(pentry))
@@ -745,7 +752,7 @@ struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
continue;
if (page_allocated) {
swap_readpage(page, false);
- if (i != swap_ra->offset &&
+ if (i != ra_info.offset &&
likely(!PageTransCompound(page))) {
SetPageReadahead(page);
count_vm_event(SWAP_RA);
@@ -757,7 +764,7 @@ struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
lru_add_drain();
skip:
return read_swap_cache_async(fentry, gfp_mask, vma, vmf->address,
- swap_ra->win == 1);
+ ra_info.win == 1);
}
#ifdef CONFIG_SYSFS
--
2.7.4
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH v2 1/2] mm:swap: clean up swap readahead
@ 2017-11-01 7:15 ` Minchan Kim
0 siblings, 0 replies; 6+ messages in thread
From: Minchan Kim @ 2017-11-01 7:15 UTC (permalink / raw)
To: Andrew Morton
Cc: linux-kernel, linux-mm, Hugh Dickins, Huang Ying, kernel-team,
Minchan Kim
When I see recent change of swap readahead, I am very unhappy
about current code structure which diverges two swap readahead
algorithm in do_swap_page. This patch is to clean it up.
Main motivation is that fault handler doesn't need to be aware of
readahead algorithms but just should call swapin_readahead.
As first step, this patch cleans up a little bit but not perfect
(I just separate for review easier) so next patch will make the goal
complete.
Signed-off-by: Minchan Kim <minchan@kernel.org>
---
include/linux/swap.h | 17 ++--------
mm/memory.c | 17 +++-------
mm/swap_state.c | 89 ++++++++++++++++++++++++++++------------------------
3 files changed, 55 insertions(+), 68 deletions(-)
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 84255b3da7c1..7c7c8b344bc9 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -427,12 +427,8 @@ extern struct page *__read_swap_cache_async(swp_entry_t, gfp_t,
bool *new_page_allocated);
extern struct page *swapin_readahead(swp_entry_t, gfp_t,
struct vm_area_struct *vma, unsigned long addr);
-
-extern struct page *swap_readahead_detect(struct vm_fault *vmf,
- struct vma_swap_readahead *swap_ra);
extern struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
- struct vm_fault *vmf,
- struct vma_swap_readahead *swap_ra);
+ struct vm_fault *vmf);
/* linux/mm/swapfile.c */
extern atomic_long_t nr_swap_pages;
@@ -551,15 +547,8 @@ static inline bool swap_use_vma_readahead(void)
return false;
}
-static inline struct page *swap_readahead_detect(
- struct vm_fault *vmf, struct vma_swap_readahead *swap_ra)
-{
- return NULL;
-}
-
-static inline struct page *do_swap_page_readahead(
- swp_entry_t fentry, gfp_t gfp_mask,
- struct vm_fault *vmf, struct vma_swap_readahead *swap_ra)
+static inline struct page *do_swap_page_readahead(swp_entry_t fentry,
+ gfp_t gfp_mask, struct vm_fault *vmf)
{
return NULL;
}
diff --git a/mm/memory.c b/mm/memory.c
index 8a0c410037d2..e955298e4290 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2849,21 +2849,14 @@ int do_swap_page(struct vm_fault *vmf)
struct vm_area_struct *vma = vmf->vma;
struct page *page = NULL, *swapcache = NULL;
struct mem_cgroup *memcg;
- struct vma_swap_readahead swap_ra;
swp_entry_t entry;
pte_t pte;
int locked;
int exclusive = 0;
int ret = 0;
- bool vma_readahead = swap_use_vma_readahead();
- if (vma_readahead)
- page = swap_readahead_detect(vmf, &swap_ra);
- if (!pte_unmap_same(vma->vm_mm, vmf->pmd, vmf->pte, vmf->orig_pte)) {
- if (page)
- put_page(page);
+ if (!pte_unmap_same(vma->vm_mm, vmf->pmd, vmf->pte, vmf->orig_pte))
goto out;
- }
entry = pte_to_swp_entry(vmf->orig_pte);
if (unlikely(non_swap_entry(entry))) {
@@ -2889,9 +2882,7 @@ int do_swap_page(struct vm_fault *vmf)
delayacct_set_flag(DELAYACCT_PF_SWAPIN);
- if (!page)
- page = lookup_swap_cache(entry, vma_readahead ? vma : NULL,
- vmf->address);
+ page = lookup_swap_cache(entry, vma, vmf->address);
if (!page) {
struct swap_info_struct *si = swp_swap_info(entry);
@@ -2907,9 +2898,9 @@ int do_swap_page(struct vm_fault *vmf)
swap_readpage(page, true);
}
} else {
- if (vma_readahead)
+ if (swap_use_vma_readahead())
page = do_swap_page_readahead(entry,
- GFP_HIGHUSER_MOVABLE, vmf, &swap_ra);
+ GFP_HIGHUSER_MOVABLE, vmf);
else
page = swapin_readahead(entry,
GFP_HIGHUSER_MOVABLE, vma, vmf->address);
diff --git a/mm/swap_state.c b/mm/swap_state.c
index 6c017ced11e6..e3c535fcd2df 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -331,32 +331,38 @@ struct page *lookup_swap_cache(swp_entry_t entry, struct vm_area_struct *vma,
unsigned long addr)
{
struct page *page;
- unsigned long ra_info;
- int win, hits, readahead;
page = find_get_page(swap_address_space(entry), swp_offset(entry));
INC_CACHE_INFO(find_total);
if (page) {
+ bool vma_ra = swap_use_vma_readahead();
+ bool readahead = TestClearPageReadahead(page);
+
INC_CACHE_INFO(find_success);
if (unlikely(PageTransCompound(page)))
return page;
- readahead = TestClearPageReadahead(page);
- if (vma) {
- ra_info = GET_SWAP_RA_VAL(vma);
- win = SWAP_RA_WIN(ra_info);
- hits = SWAP_RA_HITS(ra_info);
+
+ if (vma && vma_ra) {
+ unsigned long ra_val;
+ int win, hits;
+
+ ra_val = GET_SWAP_RA_VAL(vma);
+ win = SWAP_RA_WIN(ra_val);
+ hits = SWAP_RA_HITS(ra_val);
if (readahead)
hits = min_t(int, hits + 1, SWAP_RA_HITS_MAX);
atomic_long_set(&vma->swap_readahead_info,
SWAP_RA_VAL(addr, win, hits));
}
+
if (readahead) {
count_vm_event(SWAP_RA_HIT);
- if (!vma)
+ if (!vma || !vma_ra)
atomic_inc(&swapin_readahead_hits);
}
}
+
return page;
}
@@ -645,16 +651,15 @@ static inline void swap_ra_clamp_pfn(struct vm_area_struct *vma,
PFN_DOWN((faddr & PMD_MASK) + PMD_SIZE));
}
-struct page *swap_readahead_detect(struct vm_fault *vmf,
- struct vma_swap_readahead *swap_ra)
+static void swap_ra_info(struct vm_fault *vmf,
+ struct vma_swap_readahead *ra_info)
{
struct vm_area_struct *vma = vmf->vma;
- unsigned long swap_ra_info;
- struct page *page;
+ unsigned long ra_val;
swp_entry_t entry;
unsigned long faddr, pfn, fpfn;
unsigned long start, end;
- pte_t *pte;
+ pte_t *pte, *orig_pte;
unsigned int max_win, hits, prev_win, win, left;
#ifndef CONFIG_64BIT
pte_t *tpte;
@@ -663,30 +668,32 @@ struct page *swap_readahead_detect(struct vm_fault *vmf,
max_win = 1 << min_t(unsigned int, READ_ONCE(page_cluster),
SWAP_RA_ORDER_CEILING);
if (max_win == 1) {
- swap_ra->win = 1;
- return NULL;
+ ra_info->win = 1;
+ return;
}
faddr = vmf->address;
- entry = pte_to_swp_entry(vmf->orig_pte);
- if ((unlikely(non_swap_entry(entry))))
- return NULL;
- page = lookup_swap_cache(entry, vma, faddr);
- if (page)
- return page;
+ orig_pte = pte = pte_offset_map(vmf->pmd, faddr);
+ entry = pte_to_swp_entry(*pte);
+ if ((unlikely(non_swap_entry(entry)))) {
+ pte_unmap(orig_pte);
+ return;
+ }
fpfn = PFN_DOWN(faddr);
- swap_ra_info = GET_SWAP_RA_VAL(vma);
- pfn = PFN_DOWN(SWAP_RA_ADDR(swap_ra_info));
- prev_win = SWAP_RA_WIN(swap_ra_info);
- hits = SWAP_RA_HITS(swap_ra_info);
- swap_ra->win = win = __swapin_nr_pages(pfn, fpfn, hits,
+ ra_val = GET_SWAP_RA_VAL(vma);
+ pfn = PFN_DOWN(SWAP_RA_ADDR(ra_val));
+ prev_win = SWAP_RA_WIN(ra_val);
+ hits = SWAP_RA_HITS(ra_val);
+ ra_info->win = win = __swapin_nr_pages(pfn, fpfn, hits,
max_win, prev_win);
atomic_long_set(&vma->swap_readahead_info,
SWAP_RA_VAL(faddr, win, 0));
- if (win == 1)
- return NULL;
+ if (win == 1) {
+ pte_unmap(orig_pte);
+ return;
+ }
/* Copy the PTEs because the page table may be unmapped */
if (fpfn == pfn + 1)
@@ -699,23 +706,21 @@ struct page *swap_readahead_detect(struct vm_fault *vmf,
swap_ra_clamp_pfn(vma, faddr, fpfn - left, fpfn + win - left,
&start, &end);
}
- swap_ra->nr_pte = end - start;
- swap_ra->offset = fpfn - start;
- pte = vmf->pte - swap_ra->offset;
+ ra_info->nr_pte = end - start;
+ ra_info->offset = fpfn - start;
+ pte -= ra_info->offset;
#ifdef CONFIG_64BIT
- swap_ra->ptes = pte;
+ ra_info->ptes = pte;
#else
- tpte = swap_ra->ptes;
+ tpte = ra_info->ptes;
for (pfn = start; pfn != end; pfn++)
*tpte++ = *pte++;
#endif
-
- return NULL;
+ pte_unmap(orig_pte);
}
struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
- struct vm_fault *vmf,
- struct vma_swap_readahead *swap_ra)
+ struct vm_fault *vmf)
{
struct blk_plug plug;
struct vm_area_struct *vma = vmf->vma;
@@ -724,12 +729,14 @@ struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
swp_entry_t entry;
unsigned int i;
bool page_allocated;
+ struct vma_swap_readahead ra_info = {0,};
- if (swap_ra->win == 1)
+ swap_ra_info(vmf, &ra_info);
+ if (ra_info.win == 1)
goto skip;
blk_start_plug(&plug);
- for (i = 0, pte = swap_ra->ptes; i < swap_ra->nr_pte;
+ for (i = 0, pte = ra_info.ptes; i < ra_info.nr_pte;
i++, pte++) {
pentry = *pte;
if (pte_none(pentry))
@@ -745,7 +752,7 @@ struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
continue;
if (page_allocated) {
swap_readpage(page, false);
- if (i != swap_ra->offset &&
+ if (i != ra_info.offset &&
likely(!PageTransCompound(page))) {
SetPageReadahead(page);
count_vm_event(SWAP_RA);
@@ -757,7 +764,7 @@ struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
lru_add_drain();
skip:
return read_swap_cache_async(fentry, gfp_mask, vma, vmf->address,
- swap_ra->win == 1);
+ ra_info.win == 1);
}
#ifdef CONFIG_SYSFS
--
2.7.4
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v2 2/2] mm:swap: unify cluster-based and vma-based swap readahead
2017-11-01 7:15 ` Minchan Kim
@ 2017-11-01 7:15 ` Minchan Kim
-1 siblings, 0 replies; 6+ messages in thread
From: Minchan Kim @ 2017-11-01 7:15 UTC (permalink / raw)
To: Andrew Morton
Cc: linux-kernel, linux-mm, Hugh Dickins, Huang Ying, kernel-team,
Minchan Kim
This patch makes do_swap_page no need to be aware of two different
swap readahead algorithm. Just unify cluster-based and vma-based
readahead function call.
Signed-off-by: Minchan Kim <minchan@kernel.org>
---
include/linux/swap.h | 27 ++++++++-------------------
mm/memory.c | 11 ++++-------
mm/shmem.c | 5 ++++-
mm/swap_state.c | 48 +++++++++++++++++++++++++++++++++++++-----------
4 files changed, 53 insertions(+), 38 deletions(-)
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 7c7c8b344bc9..95fca979b1c1 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -403,7 +403,6 @@ int generic_swapfile_activate(struct swap_info_struct *, struct file *,
#define SWAP_ADDRESS_SPACE_SHIFT 14
#define SWAP_ADDRESS_SPACE_PAGES (1 << SWAP_ADDRESS_SPACE_SHIFT)
extern struct address_space *swapper_spaces[];
-extern bool swap_vma_readahead;
#define swap_address_space(entry) \
(&swapper_spaces[swp_type(entry)][swp_offset(entry) \
>> SWAP_ADDRESS_SPACE_SHIFT])
@@ -425,10 +424,10 @@ extern struct page *read_swap_cache_async(swp_entry_t, gfp_t,
extern struct page *__read_swap_cache_async(swp_entry_t, gfp_t,
struct vm_area_struct *vma, unsigned long addr,
bool *new_page_allocated);
-extern struct page *swapin_readahead(swp_entry_t, gfp_t,
- struct vm_area_struct *vma, unsigned long addr);
-extern struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
- struct vm_fault *vmf);
+extern struct page *swap_cluster_readahead(swp_entry_t entry, gfp_t flag,
+ struct vm_fault *vmf);
+extern struct page *swapin_readahead(swp_entry_t entry, gfp_t flag,
+ struct vm_fault *vmf);
/* linux/mm/swapfile.c */
extern atomic_long_t nr_swap_pages;
@@ -436,11 +435,6 @@ extern long total_swap_pages;
extern atomic_t nr_rotate_swap;
extern bool has_usable_swap(void);
-static inline bool swap_use_vma_readahead(void)
-{
- return READ_ONCE(swap_vma_readahead) && !atomic_read(&nr_rotate_swap);
-}
-
/* Swap 50% full? Release swapcache more aggressively.. */
static inline bool vm_swap_full(void)
{
@@ -536,19 +530,14 @@ static inline void put_swap_page(struct page *page, swp_entry_t swp)
{
}
-static inline struct page *swapin_readahead(swp_entry_t swp, gfp_t gfp_mask,
- struct vm_area_struct *vma, unsigned long addr)
+static inline struct page *swap_cluster_readahead(swp_entry_t entry,
+ gfp_t gfp_mask, struct vm_fault *vmf)
{
return NULL;
}
-static inline bool swap_use_vma_readahead(void)
-{
- return false;
-}
-
-static inline struct page *do_swap_page_readahead(swp_entry_t fentry,
- gfp_t gfp_mask, struct vm_fault *vmf)
+static inline struct page *swapin_readahead(swp_entry_t swp, gfp_t gfp_mask,
+ struct vm_fault *vmf)
{
return NULL;
}
diff --git a/mm/memory.c b/mm/memory.c
index e955298e4290..ce5e3d7ccc5c 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2889,7 +2889,8 @@ int do_swap_page(struct vm_fault *vmf)
if (si->flags & SWP_SYNCHRONOUS_IO &&
__swap_count(si, entry) == 1) {
/* skip swapcache */
- page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, vmf->address);
+ page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma,
+ vmf->address);
if (page) {
__SetPageLocked(page);
__SetPageSwapBacked(page);
@@ -2898,12 +2899,8 @@ int do_swap_page(struct vm_fault *vmf)
swap_readpage(page, true);
}
} else {
- if (swap_use_vma_readahead())
- page = do_swap_page_readahead(entry,
- GFP_HIGHUSER_MOVABLE, vmf);
- else
- page = swapin_readahead(entry,
- GFP_HIGHUSER_MOVABLE, vma, vmf->address);
+ page = swapin_readahead(entry, GFP_HIGHUSER_MOVABLE,
+ vmf);
swapcache = page;
}
diff --git a/mm/shmem.c b/mm/shmem.c
index 62dfdc097e44..d92923834d68 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1413,9 +1413,12 @@ static struct page *shmem_swapin(swp_entry_t swap, gfp_t gfp,
{
struct vm_area_struct pvma;
struct page *page;
+ struct vm_fault vmf;
shmem_pseudo_vma_init(&pvma, info, index);
- page = swapin_readahead(swap, gfp, &pvma, 0);
+ vmf.vma = &pvma;
+ vmf.address = 0;
+ page = swap_cluster_readahead(swap, gfp, &vmf);
shmem_pseudo_vma_destroy(&pvma);
return page;
diff --git a/mm/swap_state.c b/mm/swap_state.c
index e3c535fcd2df..94e01413eb43 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -37,7 +37,7 @@ static const struct address_space_operations swap_aops = {
struct address_space *swapper_spaces[MAX_SWAPFILES] __read_mostly;
static unsigned int nr_swapper_spaces[MAX_SWAPFILES] __read_mostly;
-bool swap_vma_readahead __read_mostly = true;
+bool enable_vma_readahead __read_mostly = true;
#define SWAP_RA_WIN_SHIFT (PAGE_SHIFT / 2)
#define SWAP_RA_HITS_MASK ((1UL << SWAP_RA_WIN_SHIFT) - 1)
@@ -321,6 +321,11 @@ void free_pages_and_swap_cache(struct page **pages, int nr)
release_pages(pagep, nr);
}
+static inline bool swap_use_vma_readahead(void)
+{
+ return READ_ONCE(enable_vma_readahead) && !atomic_read(&nr_rotate_swap);
+}
+
/*
* Lookup a swap entry in the swap cache. A found page will be returned
* unlocked and with its refcount incremented - we rely on the kernel
@@ -538,11 +543,10 @@ static unsigned long swapin_nr_pages(unsigned long offset)
}
/**
- * swapin_readahead - swap in pages in hope we need them soon
+ * swap_cluster_readahead - swap in pages in hope we need them soon
* @entry: swap entry of this memory
* @gfp_mask: memory allocation flags
- * @vma: user vma this address belongs to
- * @addr: target address for mempolicy
+ * @vmf: fault information
*
* Returns the struct page for entry and addr, after queueing swapin.
*
@@ -554,10 +558,10 @@ static unsigned long swapin_nr_pages(unsigned long offset)
* This has been extended to use the NUMA policies from the mm triggering
* the readahead.
*
- * Caller must hold down_read on the vma->vm_mm if vma is not NULL.
+ * Caller must hold down_read on the vma->vm_mm if vmf->vma is not NULL.
*/
-struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
- struct vm_area_struct *vma, unsigned long addr)
+struct page *swap_cluster_readahead(swp_entry_t entry, gfp_t gfp_mask,
+ struct vm_fault *vmf)
{
struct page *page;
unsigned long entry_offset = swp_offset(entry);
@@ -566,6 +570,8 @@ struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
unsigned long mask;
struct blk_plug plug;
bool do_poll = true, page_allocated;
+ struct vm_area_struct *vma = vmf->vma;
+ unsigned long addr = vmf->address;
mask = swapin_nr_pages(offset) - 1;
if (!mask)
@@ -719,7 +725,7 @@ static void swap_ra_info(struct vm_fault *vmf,
pte_unmap(orig_pte);
}
-struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
+struct page *swap_vma_readahead(swp_entry_t fentry, gfp_t gfp_mask,
struct vm_fault *vmf)
{
struct blk_plug plug;
@@ -767,20 +773,40 @@ struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
ra_info.win == 1);
}
+/**
+ * swapin_readahead - swap in pages in hope we need them soon
+ * @entry: swap entry of this memory
+ * @gfp_mask: memory allocation flags
+ * @vmf: fault information
+ *
+ * Returns the struct page for entry and addr, after queueing swapin.
+ *
+ * It's a main entry function for swap readahead. By the configuration,
+ * it will read ahead blocks by cluster-based(ie, physical disk based)
+ * or vma-based(ie, virtual address based on faulty address) readahead.
+ */
+struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
+ struct vm_fault *vmf)
+{
+ return swap_use_vma_readahead() ?
+ swap_vma_readahead(entry, gfp_mask, vmf) :
+ swap_cluster_readahead(entry, gfp_mask, vmf);
+}
+
#ifdef CONFIG_SYSFS
static ssize_t vma_ra_enabled_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
- return sprintf(buf, "%s\n", swap_vma_readahead ? "true" : "false");
+ return sprintf(buf, "%s\n", enable_vma_readahead ? "true" : "false");
}
static ssize_t vma_ra_enabled_store(struct kobject *kobj,
struct kobj_attribute *attr,
const char *buf, size_t count)
{
if (!strncmp(buf, "true", 4) || !strncmp(buf, "1", 1))
- swap_vma_readahead = true;
+ enable_vma_readahead = true;
else if (!strncmp(buf, "false", 5) || !strncmp(buf, "0", 1))
- swap_vma_readahead = false;
+ enable_vma_readahead = false;
else
return -EINVAL;
--
2.7.4
^ permalink raw reply related [flat|nested] 6+ messages in thread* [PATCH v2 2/2] mm:swap: unify cluster-based and vma-based swap readahead
@ 2017-11-01 7:15 ` Minchan Kim
0 siblings, 0 replies; 6+ messages in thread
From: Minchan Kim @ 2017-11-01 7:15 UTC (permalink / raw)
To: Andrew Morton
Cc: linux-kernel, linux-mm, Hugh Dickins, Huang Ying, kernel-team,
Minchan Kim
This patch makes do_swap_page no need to be aware of two different
swap readahead algorithm. Just unify cluster-based and vma-based
readahead function call.
Signed-off-by: Minchan Kim <minchan@kernel.org>
---
include/linux/swap.h | 27 ++++++++-------------------
mm/memory.c | 11 ++++-------
mm/shmem.c | 5 ++++-
mm/swap_state.c | 48 +++++++++++++++++++++++++++++++++++++-----------
4 files changed, 53 insertions(+), 38 deletions(-)
diff --git a/include/linux/swap.h b/include/linux/swap.h
index 7c7c8b344bc9..95fca979b1c1 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -403,7 +403,6 @@ int generic_swapfile_activate(struct swap_info_struct *, struct file *,
#define SWAP_ADDRESS_SPACE_SHIFT 14
#define SWAP_ADDRESS_SPACE_PAGES (1 << SWAP_ADDRESS_SPACE_SHIFT)
extern struct address_space *swapper_spaces[];
-extern bool swap_vma_readahead;
#define swap_address_space(entry) \
(&swapper_spaces[swp_type(entry)][swp_offset(entry) \
>> SWAP_ADDRESS_SPACE_SHIFT])
@@ -425,10 +424,10 @@ extern struct page *read_swap_cache_async(swp_entry_t, gfp_t,
extern struct page *__read_swap_cache_async(swp_entry_t, gfp_t,
struct vm_area_struct *vma, unsigned long addr,
bool *new_page_allocated);
-extern struct page *swapin_readahead(swp_entry_t, gfp_t,
- struct vm_area_struct *vma, unsigned long addr);
-extern struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
- struct vm_fault *vmf);
+extern struct page *swap_cluster_readahead(swp_entry_t entry, gfp_t flag,
+ struct vm_fault *vmf);
+extern struct page *swapin_readahead(swp_entry_t entry, gfp_t flag,
+ struct vm_fault *vmf);
/* linux/mm/swapfile.c */
extern atomic_long_t nr_swap_pages;
@@ -436,11 +435,6 @@ extern long total_swap_pages;
extern atomic_t nr_rotate_swap;
extern bool has_usable_swap(void);
-static inline bool swap_use_vma_readahead(void)
-{
- return READ_ONCE(swap_vma_readahead) && !atomic_read(&nr_rotate_swap);
-}
-
/* Swap 50% full? Release swapcache more aggressively.. */
static inline bool vm_swap_full(void)
{
@@ -536,19 +530,14 @@ static inline void put_swap_page(struct page *page, swp_entry_t swp)
{
}
-static inline struct page *swapin_readahead(swp_entry_t swp, gfp_t gfp_mask,
- struct vm_area_struct *vma, unsigned long addr)
+static inline struct page *swap_cluster_readahead(swp_entry_t entry,
+ gfp_t gfp_mask, struct vm_fault *vmf)
{
return NULL;
}
-static inline bool swap_use_vma_readahead(void)
-{
- return false;
-}
-
-static inline struct page *do_swap_page_readahead(swp_entry_t fentry,
- gfp_t gfp_mask, struct vm_fault *vmf)
+static inline struct page *swapin_readahead(swp_entry_t swp, gfp_t gfp_mask,
+ struct vm_fault *vmf)
{
return NULL;
}
diff --git a/mm/memory.c b/mm/memory.c
index e955298e4290..ce5e3d7ccc5c 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2889,7 +2889,8 @@ int do_swap_page(struct vm_fault *vmf)
if (si->flags & SWP_SYNCHRONOUS_IO &&
__swap_count(si, entry) == 1) {
/* skip swapcache */
- page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, vmf->address);
+ page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma,
+ vmf->address);
if (page) {
__SetPageLocked(page);
__SetPageSwapBacked(page);
@@ -2898,12 +2899,8 @@ int do_swap_page(struct vm_fault *vmf)
swap_readpage(page, true);
}
} else {
- if (swap_use_vma_readahead())
- page = do_swap_page_readahead(entry,
- GFP_HIGHUSER_MOVABLE, vmf);
- else
- page = swapin_readahead(entry,
- GFP_HIGHUSER_MOVABLE, vma, vmf->address);
+ page = swapin_readahead(entry, GFP_HIGHUSER_MOVABLE,
+ vmf);
swapcache = page;
}
diff --git a/mm/shmem.c b/mm/shmem.c
index 62dfdc097e44..d92923834d68 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1413,9 +1413,12 @@ static struct page *shmem_swapin(swp_entry_t swap, gfp_t gfp,
{
struct vm_area_struct pvma;
struct page *page;
+ struct vm_fault vmf;
shmem_pseudo_vma_init(&pvma, info, index);
- page = swapin_readahead(swap, gfp, &pvma, 0);
+ vmf.vma = &pvma;
+ vmf.address = 0;
+ page = swap_cluster_readahead(swap, gfp, &vmf);
shmem_pseudo_vma_destroy(&pvma);
return page;
diff --git a/mm/swap_state.c b/mm/swap_state.c
index e3c535fcd2df..94e01413eb43 100644
--- a/mm/swap_state.c
+++ b/mm/swap_state.c
@@ -37,7 +37,7 @@ static const struct address_space_operations swap_aops = {
struct address_space *swapper_spaces[MAX_SWAPFILES] __read_mostly;
static unsigned int nr_swapper_spaces[MAX_SWAPFILES] __read_mostly;
-bool swap_vma_readahead __read_mostly = true;
+bool enable_vma_readahead __read_mostly = true;
#define SWAP_RA_WIN_SHIFT (PAGE_SHIFT / 2)
#define SWAP_RA_HITS_MASK ((1UL << SWAP_RA_WIN_SHIFT) - 1)
@@ -321,6 +321,11 @@ void free_pages_and_swap_cache(struct page **pages, int nr)
release_pages(pagep, nr);
}
+static inline bool swap_use_vma_readahead(void)
+{
+ return READ_ONCE(enable_vma_readahead) && !atomic_read(&nr_rotate_swap);
+}
+
/*
* Lookup a swap entry in the swap cache. A found page will be returned
* unlocked and with its refcount incremented - we rely on the kernel
@@ -538,11 +543,10 @@ static unsigned long swapin_nr_pages(unsigned long offset)
}
/**
- * swapin_readahead - swap in pages in hope we need them soon
+ * swap_cluster_readahead - swap in pages in hope we need them soon
* @entry: swap entry of this memory
* @gfp_mask: memory allocation flags
- * @vma: user vma this address belongs to
- * @addr: target address for mempolicy
+ * @vmf: fault information
*
* Returns the struct page for entry and addr, after queueing swapin.
*
@@ -554,10 +558,10 @@ static unsigned long swapin_nr_pages(unsigned long offset)
* This has been extended to use the NUMA policies from the mm triggering
* the readahead.
*
- * Caller must hold down_read on the vma->vm_mm if vma is not NULL.
+ * Caller must hold down_read on the vma->vm_mm if vmf->vma is not NULL.
*/
-struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
- struct vm_area_struct *vma, unsigned long addr)
+struct page *swap_cluster_readahead(swp_entry_t entry, gfp_t gfp_mask,
+ struct vm_fault *vmf)
{
struct page *page;
unsigned long entry_offset = swp_offset(entry);
@@ -566,6 +570,8 @@ struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
unsigned long mask;
struct blk_plug plug;
bool do_poll = true, page_allocated;
+ struct vm_area_struct *vma = vmf->vma;
+ unsigned long addr = vmf->address;
mask = swapin_nr_pages(offset) - 1;
if (!mask)
@@ -719,7 +725,7 @@ static void swap_ra_info(struct vm_fault *vmf,
pte_unmap(orig_pte);
}
-struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
+struct page *swap_vma_readahead(swp_entry_t fentry, gfp_t gfp_mask,
struct vm_fault *vmf)
{
struct blk_plug plug;
@@ -767,20 +773,40 @@ struct page *do_swap_page_readahead(swp_entry_t fentry, gfp_t gfp_mask,
ra_info.win == 1);
}
+/**
+ * swapin_readahead - swap in pages in hope we need them soon
+ * @entry: swap entry of this memory
+ * @gfp_mask: memory allocation flags
+ * @vmf: fault information
+ *
+ * Returns the struct page for entry and addr, after queueing swapin.
+ *
+ * It's a main entry function for swap readahead. By the configuration,
+ * it will read ahead blocks by cluster-based(ie, physical disk based)
+ * or vma-based(ie, virtual address based on faulty address) readahead.
+ */
+struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask,
+ struct vm_fault *vmf)
+{
+ return swap_use_vma_readahead() ?
+ swap_vma_readahead(entry, gfp_mask, vmf) :
+ swap_cluster_readahead(entry, gfp_mask, vmf);
+}
+
#ifdef CONFIG_SYSFS
static ssize_t vma_ra_enabled_show(struct kobject *kobj,
struct kobj_attribute *attr, char *buf)
{
- return sprintf(buf, "%s\n", swap_vma_readahead ? "true" : "false");
+ return sprintf(buf, "%s\n", enable_vma_readahead ? "true" : "false");
}
static ssize_t vma_ra_enabled_store(struct kobject *kobj,
struct kobj_attribute *attr,
const char *buf, size_t count)
{
if (!strncmp(buf, "true", 4) || !strncmp(buf, "1", 1))
- swap_vma_readahead = true;
+ enable_vma_readahead = true;
else if (!strncmp(buf, "false", 5) || !strncmp(buf, "0", 1))
- swap_vma_readahead = false;
+ enable_vma_readahead = false;
else
return -EINVAL;
--
2.7.4
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related [flat|nested] 6+ messages in thread