Skip to contentSkip to conversation

vduse: avoid using __GFP_NOFAIL

PATCH In review Jason Wang 2 participants 3 messages Aug 8, 2024 – Aug 16, 2024

Subsystems (2)
lore
Share & tools
lore

Jason Wang

author
#0 lore raw

PATCH v2 1/1 vduse: avoid using __GFP_NOFAIL

Barry said [1]:

"""
mm doesn't support non-blockable __GFP_NOFAIL allocation. Because __GFP_NOFAIL without direct reclamation may just result in a busy loop within non-sleepable contexts.

The current code will result in returning a NULL pointer but not a busy-loop.

static inline struct page *
__alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
struct alloc_context *ac)

{
        ...
        /*
         * Make sure that __GFP_NOFAIL request doesn't leak out and make sure
         * we always retry
         */
        if (gfp_mask & __GFP_NOFAIL) {
                /*
                 * All existing users of the __GFP_NOFAIL are blockable, so warn
                 * of any new users that actually require GFP_NOWAIT
                 */
                if (WARN_ON_ONCE_GFP(!can_direct_reclaim, gfp_mask))
                        goto fail;
                ...
        }
        ...
fail:
        warn_alloc(gfp_mask, ac->nodemask,
                        "page allocation failure: order:%u", order);
got_pg:
        return page;
}

We have two choices to address the issue:
1. busy-loop
2. BUG_ON

the below patch chose 2:
https://lore.kernel.org/linux-mm/[email protected]/
""“

Unfortuantely, we do that under read lock. A possible way to fix that is to move the pages allocation out of the lock into the caller, but having to allocate a huge number of pages and auxiliary page array seems to be problematic as well per Tetsuon [2]:

"""
You should implement proper error handling instead of using __GFP_NOFAIL if count can become large.
"""

So I choose another way, which does not release kernel bounce pages when user tries to register usersapce bounce pages. Then we don't need to do allocation in the path which is not expected to be fail (e.g in the release). We pay this for more memory usage as we don't release kernel bounce pages but further optimizations could be done on top.

[1] https://lore.kernel.org/all/CACGkMEtcOJAA96SF9B8m-nZ1X04-XZr+nq8ZQ2saLnUdfOGOLg@mail.gmail.com/T/#m3caef86a66ea6318ef94f9976ddb3a0ccfe6fcf8
[2] https://lore.kernel.org/all/CACGkMEtcOJAA96SF9B8m-nZ1X04-XZr+nq8ZQ2saLnUdfOGOLg@mail.gmail.com/T/#m7ad10eaba48ade5abf2d572f24e185d9fb146480

Fixes 6c77ed22880d ("vduse: Support using userspace pages as bounce buffer")

---
Changes since V1:
- Tweak the commit log
- Assign map->user_bounce_page to NULL for safety

---
 drivers/vdpa/vdpa_user/iova_domain.c | 19 +++++++++++--------
 drivers/vdpa/vdpa_user/iova_domain.h |  1 +
 2 files changed, 12 insertions(+), 8 deletions(-)
drivers/vdpa/vdpa_user/iova_domain.c +11 -8
index 791d38d6284c..58116f89d8da 100644
@@ -162,6 +162,7 @@ static void vduse_domain_bounce(struct vduse_iova_domain *domain,
 				enum dma_data_direction dir)
 {
 	struct vduse_bounce_map *map;
+	struct page *page;
 	unsigned int offset;
 	void *addr;
 	size_t sz;
@@ -178,7 +179,10 @@ static void vduse_domain_bounce(struct vduse_iova_domain *domain,
 			    map->orig_phys == INVALID_PHYS_ADDR))
 			return;
 
-		addr = kmap_local_page(map->bounce_page);
+		page = domain->user_bounce_pages ?
+		       map->user_bounce_page : map->bounce_page;
+
+		addr = kmap_local_page(page);
 		do_bounce(map->orig_phys + offset, addr + offset, sz, dir);
 		kunmap_local(addr);
 		size -= sz;
@@ -270,9 +274,8 @@ int vduse_domain_add_user_bounce_pages(struct vduse_iova_domain *domain,
 				memcpy_to_page(pages[i], 0,
 					       page_address(map->bounce_page),
 					       PAGE_SIZE);
-			__free_page(map->bounce_page);
 		}
-		map->bounce_page = pages[i];
+		map->user_bounce_page = pages[i];
 		get_page(pages[i]);
 	}
 	domain->user_bounce_pages = true;
@@ -297,17 +300,17 @@ void vduse_domain_remove_user_bounce_pages(struct vduse_iova_domain *domain)
 		struct page *page = NULL;
 
 		map = &domain->bounce_maps[i];
-		if (WARN_ON(!map->bounce_page))
+		if (WARN_ON(!map->user_bounce_page))
 			continue;
 
 		/* Copy user page to kernel page if it's in use */
 		if (map->orig_phys != INVALID_PHYS_ADDR) {
-			page = alloc_page(GFP_ATOMIC | __GFP_NOFAIL);
+			page = map->bounce_page;
 			memcpy_from_page(page_address(page),
-					 map->bounce_page, 0, PAGE_SIZE);
+					 map->user_bounce_page, 0, PAGE_SIZE);
 		}
-		put_page(map->bounce_page);
-		map->bounce_page = page;
+		put_page(map->user_bounce_page);
+		map->user_bounce_page = NULL;
 	}
 	domain->user_bounce_pages = false;
 out:
drivers/vdpa/vdpa_user/iova_domain.h +1 -0
index f92f22a7267d..7f3f0928ec78 100644
@@ -21,6 +21,7 @@
 
 struct vduse_bounce_map {
 	struct page *bounce_page;
+	struct page *user_bounce_page;
 	u64 orig_phys;
 };
 
Signature · 1 line
2.31.1
#1 lore raw
Quoted 61 lines · Jason Wang

Barry said [1]:

"""
mm doesn't support non-blockable __GFP_NOFAIL allocation. Because __GFP_NOFAIL without direct reclamation may just result in a busy loop within non-sleepable contexts.

The current code will result in returning a NULL pointer but not a busy-loop.

static inline struct page *
__alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
struct alloc_context *ac)

{
        ...
        /*
         * Make sure that __GFP_NOFAIL request doesn't leak out and make sure
         * we always retry
         */
        if (gfp_mask & __GFP_NOFAIL) {
                /*
                 * All existing users of the __GFP_NOFAIL are blockable, so warn
                 * of any new users that actually require GFP_NOWAIT
                 */
                if (WARN_ON_ONCE_GFP(!can_direct_reclaim, gfp_mask))
                        goto fail;
                ...
        }
        ...
fail:
        warn_alloc(gfp_mask, ac->nodemask,
                        "page allocation failure: order:%u", order);
got_pg:
        return page;
}

We have two choices to address the issue:
1. busy-loop
2. BUG_ON

the below patch chose 2:
https://lore.kernel.org/linux-mm/[email protected]/
""“

Unfortuantely, we do that under read lock. A possible way to fix that is to move the pages allocation out of the lock into the caller, but having to allocate a huge number of pages and auxiliary page array seems to be problematic as well per Tetsuon [2]:

"""
You should implement proper error handling instead of using __GFP_NOFAIL if count can become large.
"""

So I choose another way, which does not release kernel bounce pages when user tries to register usersapce bounce pages. Then we don't need to do allocation in the path which is not expected to be fail (e.g in the release). We pay this for more memory usage as we don't release kernel bounce pages but further optimizations could be done on top.

Quoted 7 lines

[1] https://lore.kernel.org/all/CACGkMEtcOJAA96SF9B8m-nZ1X04-XZr+nq8ZQ2saLnUdfOGOLg@mail.gmail.com/T/#m3caef86a66ea6318ef94f9976ddb3a0ccfe6fcf8
[2] https://lore.kernel.org/all/CACGkMEtcOJAA96SF9B8m-nZ1X04-XZr+nq8ZQ2saLnUdfOGOLg@mail.gmail.com/T/#m7ad10eaba48ade5abf2d572f24e185d9fb146480

Fixes 6c77ed22880d ("vduse: Support using userspace pages as bounce buffer")

Hi Jason,

Is this the final version to fix __GFP_NOFAIL issue, or will you have a new version?
Do you prefer it to go through the mm tree or the drivers subsystem tree? If it's the former, I can send a new version that includes this one in my series[1].

[1] https://lore.kernel.org/linux-mm/[email protected]/

Quoted 80 lines

---
Changes since V1:
- Tweak the commit log
- Assign map->user_bounce_page to NULL for safety

---
 drivers/vdpa/vdpa_user/iova_domain.c | 19 +++++++++++--------
 drivers/vdpa/vdpa_user/iova_domain.h |  1 +
 2 files changed, 12 insertions(+), 8 deletions(-)
drivers/vdpa/vdpa_user/iova_domain.c +11 -8
index 791d38d6284c..58116f89d8da 100644
@@ -162,6 +162,7 @@ static void vduse_domain_bounce(struct vduse_iova_domain *domain,
                                enum dma_data_direction dir)
 {
        struct vduse_bounce_map *map;
+       struct page *page;
        unsigned int offset;
        void *addr;
        size_t sz;
@@ -178,7 +179,10 @@ static void vduse_domain_bounce(struct vduse_iova_domain *domain,
                            map->orig_phys == INVALID_PHYS_ADDR))
                        return;
 
-               addr = kmap_local_page(map->bounce_page);
+               page = domain->user_bounce_pages ?
+                      map->user_bounce_page : map->bounce_page;
+
+               addr = kmap_local_page(page);
                do_bounce(map->orig_phys + offset, addr + offset, sz, dir);
                kunmap_local(addr);
                size -= sz;
@@ -270,9 +274,8 @@ int vduse_domain_add_user_bounce_pages(struct vduse_iova_domain *domain,
                                memcpy_to_page(pages[i], 0,
                                               page_address(map->bounce_page),
                                               PAGE_SIZE);
-                       __free_page(map->bounce_page);
                }
-               map->bounce_page = pages[i];
+               map->user_bounce_page = pages[i];
                get_page(pages[i]);
        }
        domain->user_bounce_pages = true;
@@ -297,17 +300,17 @@ void vduse_domain_remove_user_bounce_pages(struct vduse_iova_domain *domain)
                struct page *page = NULL;
 
                map = &domain->bounce_maps[i];
-               if (WARN_ON(!map->bounce_page))
+               if (WARN_ON(!map->user_bounce_page))
                        continue;
 
                /* Copy user page to kernel page if it's in use */
                if (map->orig_phys != INVALID_PHYS_ADDR) {
-                       page = alloc_page(GFP_ATOMIC | __GFP_NOFAIL);
+                       page = map->bounce_page;
                        memcpy_from_page(page_address(page),
-                                        map->bounce_page, 0, PAGE_SIZE);
+                                        map->user_bounce_page, 0, PAGE_SIZE);
                }
-               put_page(map->bounce_page);
-               map->bounce_page = page;
+               put_page(map->user_bounce_page);
+               map->user_bounce_page = NULL;
        }
        domain->user_bounce_pages = false;
 out:
drivers/vdpa/vdpa_user/iova_domain.h +1 -1
index f92f22a7267d..7f3f0928ec78 100644
@@ -21,6 +21,7 @@
 
 struct vduse_bounce_map {
        struct page *bounce_page;
+       struct page *user_bounce_page;
        u64 orig_phys;
 };
 
--
Quoted 2 lines

2.31.1

Thanks
Barry

author
#2 lore raw
Quoted 71 lines · Barry Song
Quoted 61 lines · Jason Wang

Barry said [1]:

"""
mm doesn't support non-blockable __GFP_NOFAIL allocation. Because __GFP_NOFAIL without direct reclamation may just result in a busy loop within non-sleepable contexts.

The current code will result in returning a NULL pointer but not a busy-loop.

static inline struct page *
__alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
struct alloc_context *ac)

{
        ...
        /*
         * Make sure that __GFP_NOFAIL request doesn't leak out and make sure
         * we always retry
         */
        if (gfp_mask & __GFP_NOFAIL) {
                /*
                 * All existing users of the __GFP_NOFAIL are blockable, so warn
                 * of any new users that actually require GFP_NOWAIT
                 */
                if (WARN_ON_ONCE_GFP(!can_direct_reclaim, gfp_mask))
                        goto fail;
                ...
        }
        ...
fail:
        warn_alloc(gfp_mask, ac->nodemask,
                        "page allocation failure: order:%u", order);
got_pg:
        return page;
}

We have two choices to address the issue:
1. busy-loop
2. BUG_ON

the below patch chose 2:
https://lore.kernel.org/linux-mm/[email protected]/
""“

Unfortuantely, we do that under read lock. A possible way to fix that is to move the pages allocation out of the lock into the caller, but having to allocate a huge number of pages and auxiliary page array seems to be problematic as well per Tetsuon [2]:

"""
You should implement proper error handling instead of using __GFP_NOFAIL if count can become large.
"""

So I choose another way, which does not release kernel bounce pages when user tries to register usersapce bounce pages. Then we don't need to do allocation in the path which is not expected to be fail (e.g in the release). We pay this for more memory usage as we don't release kernel bounce pages but further optimizations could be done on top.

Quoted 7 lines

[1] https://lore.kernel.org/all/CACGkMEtcOJAA96SF9B8m-nZ1X04-XZr+nq8ZQ2saLnUdfOGOLg@mail.gmail.com/T/#m3caef86a66ea6318ef94f9976ddb3a0ccfe6fcf8
[2] https://lore.kernel.org/all/CACGkMEtcOJAA96SF9B8m-nZ1X04-XZr+nq8ZQ2saLnUdfOGOLg@mail.gmail.com/T/#m7ad10eaba48ade5abf2d572f24e185d9fb146480

Fixes 6c77ed22880d ("vduse: Support using userspace pages as bounce buffer")
Quoted 4 lines

Hi Jason,

Is this the final version to fix __GFP_NOFAIL issue, or will you have a new version?

I think so.

Quoted 1 line

Do you prefer it to go through the mm tree or the drivers subsystem tree?

Either is fine.

Quoted 4 lines

If it's the former, I can send a new version that includes this one in my series[1].

[1] https://lore.kernel.org/linux-mm/[email protected]/

That's fine.

Thanks

Quoted 87 lines
Quoted 81 lines

---
Changes since V1:
- Tweak the commit log
- Assign map->user_bounce_page to NULL for safety

---
 drivers/vdpa/vdpa_user/iova_domain.c | 19 +++++++++++--------
 drivers/vdpa/vdpa_user/iova_domain.h |  1 +
 2 files changed, 12 insertions(+), 8 deletions(-)
drivers/vdpa/vdpa_user/iova_domain.c +11 -8
index 791d38d6284c..58116f89d8da 100644
@@ -162,6 +162,7 @@ static void vduse_domain_bounce(struct vduse_iova_domain *domain,
                                enum dma_data_direction dir)
 {
        struct vduse_bounce_map *map;
+       struct page *page;
        unsigned int offset;
        void *addr;
        size_t sz;
@@ -178,7 +179,10 @@ static void vduse_domain_bounce(struct vduse_iova_domain *domain,
                            map->orig_phys == INVALID_PHYS_ADDR))
                        return;
 
-               addr = kmap_local_page(map->bounce_page);
+               page = domain->user_bounce_pages ?
+                      map->user_bounce_page : map->bounce_page;
+
+               addr = kmap_local_page(page);
                do_bounce(map->orig_phys + offset, addr + offset, sz, dir);
                kunmap_local(addr);
                size -= sz;
@@ -270,9 +274,8 @@ int vduse_domain_add_user_bounce_pages(struct vduse_iova_domain *domain,
                                memcpy_to_page(pages[i], 0,
                                               page_address(map->bounce_page),
                                               PAGE_SIZE);
-                       __free_page(map->bounce_page);
                }
-               map->bounce_page = pages[i];
+               map->user_bounce_page = pages[i];
                get_page(pages[i]);
        }
        domain->user_bounce_pages = true;
@@ -297,17 +300,17 @@ void vduse_domain_remove_user_bounce_pages(struct vduse_iova_domain *domain)
                struct page *page = NULL;
 
                map = &domain->bounce_maps[i];
-               if (WARN_ON(!map->bounce_page))
+               if (WARN_ON(!map->user_bounce_page))
                        continue;
 
                /* Copy user page to kernel page if it's in use */
                if (map->orig_phys != INVALID_PHYS_ADDR) {
-                       page = alloc_page(GFP_ATOMIC | __GFP_NOFAIL);
+                       page = map->bounce_page;
                        memcpy_from_page(page_address(page),
-                                        map->bounce_page, 0, PAGE_SIZE);
+                                        map->user_bounce_page, 0, PAGE_SIZE);
                }
-               put_page(map->bounce_page);
-               map->bounce_page = page;
+               put_page(map->user_bounce_page);
+               map->user_bounce_page = NULL;
        }
        domain->user_bounce_pages = false;
 out:
drivers/vdpa/vdpa_user/iova_domain.h +1 -1
index f92f22a7267d..7f3f0928ec78 100644
@@ -21,6 +21,7 @@
 
 struct vduse_bounce_map {
        struct page *bounce_page;
+       struct page *user_bounce_page;
        u64 orig_phys;
 };
 
--
Quoted 2 lines

2.31.1

Thanks
Barry