[PATCH AUTOSEL 4.20 34/60] kasan, slub: move kasan_poison_slab hook before page_address

[PATCH AUTOSEL 4.20 34/60] kasan, slub: move kasan_poison_slab hook before page_address

[Sasha Levin]

From: Andrey Konovalov <andreyknvl@google.com>

[ Upstream commit a71012242837fe5e67d8c999cfc357174ed5dba0 ]

With tag based KASAN page_address() looks at the page flags to see whether
the resulting pointer needs to have a tag set.  Since we don't want to set
a tag when page_address() is called on SLAB pages, we call
page_kasan_tag_reset() in kasan_poison_slab().  However in allocate_slab()
page_address() is called before kasan_poison_slab().  Fix it by changing
the order.

[andreyknvl@google.com: fix compilation error when CONFIG_SLUB_DEBUG=n]
  Link: http://lkml.kernel.org/r/ac27cc0bbaeb414ed77bcd6671a877cf3546d56e.1550066133.git.andreyknvl@google.com
Link: http://lkml.kernel.org/r/cd895d627465a3f1c712647072d17f10883be2a1.1549921721.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Evgeniy Stepanov <eugenis@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Kostya Serebryany <kcc@google.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Qian Cai <cai@lca.pw>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 mm/slub.c | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

diff --git a/mm/slub.c b/mm/slub.c
index e3629cd7aff1..d1e053d48f46 100644
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -1075,6 +1075,16 @@ static void setup_object_debug(struct kmem_cache *s, struct page *page,
 	init_tracking(s, object);
 }
 
+static void setup_page_debug(struct kmem_cache *s, void *addr, int order)
+{
+	if (!(s->flags & SLAB_POISON))
+		return;
+
+	metadata_access_enable();
+	memset(addr, POISON_INUSE, PAGE_SIZE << order);
+	metadata_access_disable();
+}
+
 static inline int alloc_consistency_checks(struct kmem_cache *s,
 					struct page *page,
 					void *object, unsigned long addr)
@@ -1330,6 +1340,8 @@ slab_flags_t kmem_cache_flags(unsigned int object_size,
 #else /* !CONFIG_SLUB_DEBUG */
 static inline void setup_object_debug(struct kmem_cache *s,
 			struct page *page, void *object) {}
+static inline void setup_page_debug(struct kmem_cache *s,
+			void *addr, int order) {}
 
 static inline int alloc_debug_processing(struct kmem_cache *s,
 	struct page *page, void *object, unsigned long addr) { return 0; }
@@ -1640,12 +1652,11 @@ static struct page *allocate_slab(struct kmem_cache *s, gfp_t flags, int node)
 	if (page_is_pfmemalloc(page))
 		SetPageSlabPfmemalloc(page);
 
-	start = page_address(page);
+	kasan_poison_slab(page);
 
-	if (unlikely(s->flags & SLAB_POISON))
-		memset(start, POISON_INUSE, PAGE_SIZE << order);
+	start = page_address(page);
 
-	kasan_poison_slab(page);
+	setup_page_debug(s, start, order);
 
 	shuffle = shuffle_freelist(s, page);
 
-- 
2.19.1

[PATCH AUTOSEL 4.20 35/60] mm: handle lru_add_drain_all for UP properly

[Sasha Levin]

From: Michal Hocko <mhocko@suse.com>

[ Upstream commit 6ea183d60c469560e7b08a83c9804299e84ec9eb ]

Since for_each_cpu(cpu, mask) added by commit 2d3854a37e8b767a
("cpumask: introduce new API, without changing anything") did not
evaluate the mask argument if NR_CPUS == 1 due to CONFIG_SMP=n,
lru_add_drain_all() is hitting WARN_ON() at __flush_work() added by
commit 4d43d395fed12463 ("workqueue: Try to catch flush_work() without
INIT_WORK().") by unconditionally calling flush_work() [1].

Workaround this issue by using CONFIG_SMP=n specific lru_add_drain_all
implementation.  There is no real need to defer the implementation to
the workqueue as the draining is going to happen on the local cpu.  So
alias lru_add_drain_all to lru_add_drain which does all the necessary
work.

[akpm@linux-foundation.org: fix various build warnings]
[1] https://lkml.kernel.org/r/18a30387-6aa5-6123-e67c-57579ecc3f38@roeck-us.net
Link: http://lkml.kernel.org/r/20190213124334.GH4525@dhcp22.suse.cz
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reported-by: Guenter Roeck <linux@roeck-us.net>
Debugged-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 mm/swap.c | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index aa483719922e..e99ef3dcdfd5 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -320,11 +320,6 @@ static inline void activate_page_drain(int cpu)
 {
 }
 
-static bool need_activate_page_drain(int cpu)
-{
-	return false;
-}
-
 void activate_page(struct page *page)
 {
 	struct zone *zone = page_zone(page);
@@ -653,13 +648,15 @@ void lru_add_drain(void)
 	put_cpu();
 }
 
+#ifdef CONFIG_SMP
+
+static DEFINE_PER_CPU(struct work_struct, lru_add_drain_work);
+
 static void lru_add_drain_per_cpu(struct work_struct *dummy)
 {
 	lru_add_drain();
 }
 
-static DEFINE_PER_CPU(struct work_struct, lru_add_drain_work);
-
 /*
  * Doesn't need any cpu hotplug locking because we do rely on per-cpu
  * kworkers being shut down before our page_alloc_cpu_dead callback is
@@ -702,6 +699,12 @@ void lru_add_drain_all(void)
 
 	mutex_unlock(&lock);
 }
+#else
+void lru_add_drain_all(void)
+{
+	lru_add_drain();
+}
+#endif
 
 /**
  * release_pages - batched put_page()
-- 
2.19.1

[PATCH AUTOSEL 4.20 37/60] tmpfs: fix link accounting when a tmpfile is linked in

[Sasha Levin]

From: "Darrick J. Wong" <darrick.wong@oracle.com>

[ Upstream commit 1062af920c07f5b54cf5060fde3339da6df0cf6b ]

tmpfs has a peculiarity of accounting hard links as if they were
separate inodes: so that when the number of inodes is limited, as it is
by default, a user cannot soak up an unlimited amount of unreclaimable
dcache memory just by repeatedly linking a file.

But when v3.11 added O_TMPFILE, and the ability to use linkat() on the
fd, we missed accommodating this new case in tmpfs: "df -i" shows that
an extra "inode" remains accounted after the file is unlinked and the fd
closed and the actual inode evicted.  If a user repeatedly links
tmpfiles into a tmpfs, the limit will be hit (ENOSPC) even after they
are deleted.

Just skip the extra reservation from shmem_link() in this case: there's
a sense in which this first link of a tmpfile is then cheaper than a
hard link of another file, but the accounting works out, and there's
still good limiting, so no need to do anything more complicated.

Link: http://lkml.kernel.org/r/alpine.LSU.2.11.1902182134370.7035@eggly.anvils
Fixes: f4e0c30c191 ("allow the temp files created by open() to be linked to")
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Hugh Dickins <hughd@google.com>
Reported-by: Matej Kupljen <matej.kupljen@gmail.com>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 mm/shmem.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/mm/shmem.c b/mm/shmem.c
index 5d07e0b1352f..7872e3b75e57 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -2852,10 +2852,14 @@ static int shmem_link(struct dentry *old_dentry, struct inode *dir, struct dentr
 	 * No ordinary (disk based) filesystem counts links as inodes;
 	 * but each new link needs a new dentry, pinning lowmem, and
 	 * tmpfs dentries cannot be pruned until they are unlinked.
+	 * But if an O_TMPFILE file is linked into the tmpfs, the
+	 * first link must skip that, to get the accounting right.
 	 */
-	ret = shmem_reserve_inode(inode->i_sb);
-	if (ret)
-		goto out;
+	if (inode->i_nlink) {
+		ret = shmem_reserve_inode(inode->i_sb);
+		if (ret)
+			goto out;
+	}
 
 	dir->i_size += BOGO_DIRENT_SIZE;
 	inode->i_ctime = dir->i_ctime = dir->i_mtime = current_time(inode);
-- 
2.19.1

Re: [PATCH AUTOSEL 4.20 37/60] tmpfs: fix link accounting when a tmpfile is linked in

[Hugh Dickins]

AUTOSEL is wrong to select this commit without also selecting
29b00e609960 ("tmpfs: fix uninitialized return value in shmem_link")
which contains the tag
Fixes: 1062af920c07 ("tmpfs: fix link accounting when a tmpfile is linked in")
Please add 29b00e609960 for those 6 trees, or else omit 1062af920c07 for now.

Thanks,
Hugh

On Wed, 13 Mar 2019, Sasha Levin wrote:

> From: "Darrick J. Wong" <darrick.wong@oracle.com>
> 
> [ Upstream commit 1062af920c07f5b54cf5060fde3339da6df0cf6b ]
> 
> tmpfs has a peculiarity of accounting hard links as if they were
> separate inodes: so that when the number of inodes is limited, as it is
> by default, a user cannot soak up an unlimited amount of unreclaimable
> dcache memory just by repeatedly linking a file.
> 
> But when v3.11 added O_TMPFILE, and the ability to use linkat() on the
> fd, we missed accommodating this new case in tmpfs: "df -i" shows that
> an extra "inode" remains accounted after the file is unlinked and the fd
> closed and the actual inode evicted.  If a user repeatedly links
> tmpfiles into a tmpfs, the limit will be hit (ENOSPC) even after they
> are deleted.
> 
> Just skip the extra reservation from shmem_link() in this case: there's
> a sense in which this first link of a tmpfile is then cheaper than a
> hard link of another file, but the accounting works out, and there's
> still good limiting, so no need to do anything more complicated.
> 
> Link: http://lkml.kernel.org/r/alpine.LSU.2.11.1902182134370.7035@eggly.anvils
> Fixes: f4e0c30c191 ("allow the temp files created by open() to be linked to")
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> Signed-off-by: Hugh Dickins <hughd@google.com>
> Reported-by: Matej Kupljen <matej.kupljen@gmail.com>
> Acked-by: Al Viro <viro@zeniv.linux.org.uk>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>  mm/shmem.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/mm/shmem.c b/mm/shmem.c
> index 5d07e0b1352f..7872e3b75e57 100644
> --- a/mm/shmem.c
> +++ b/mm/shmem.c
> @@ -2852,10 +2852,14 @@ static int shmem_link(struct dentry *old_dentry, struct inode *dir, struct dentr
>  	 * No ordinary (disk based) filesystem counts links as inodes;
>  	 * but each new link needs a new dentry, pinning lowmem, and
>  	 * tmpfs dentries cannot be pruned until they are unlinked.
> +	 * But if an O_TMPFILE file is linked into the tmpfs, the
> +	 * first link must skip that, to get the accounting right.
>  	 */
> -	ret = shmem_reserve_inode(inode->i_sb);
> -	if (ret)
> -		goto out;
> +	if (inode->i_nlink) {
> +		ret = shmem_reserve_inode(inode->i_sb);
> +		if (ret)
> +			goto out;
> +	}
>  
>  	dir->i_size += BOGO_DIRENT_SIZE;
>  	inode->i_ctime = dir->i_ctime = dir->i_mtime = current_time(inode);
> -- 
> 2.19.1

Re: [PATCH AUTOSEL 4.20 37/60] tmpfs: fix link accounting when a tmpfile is linked in

[Sasha Levin]

On Wed, Mar 13, 2019 at 12:58:26PM -0700, Hugh Dickins wrote:
>AUTOSEL is wrong to select this commit without also selecting
>29b00e609960 ("tmpfs: fix uninitialized return value in shmem_link")
>which contains the tag
>Fixes: 1062af920c07 ("tmpfs: fix link accounting when a tmpfile is linked in")
>Please add 29b00e609960 for those 6 trees, or else omit 1062af920c07 for now.

I usually look up relevant "Fixes" tags right before I queue patches up
to avoid missing things that came up recently.

I've queued 29b00e609960 on top for all trees, thank you!

--
Thanks,
Sasha

[PATCH AUTOSEL 4.20 38/60] kasan, slab: fix conflicts with CONFIG_HARDENED_USERCOPY

[Sasha Levin]

From: Andrey Konovalov <andreyknvl@google.com>

[ Upstream commit 219667c23c68eb3dbc0d5662b9246f28477fe529 ]

Similarly to commit 96fedce27e13 ("kasan: make tag based mode work with
CONFIG_HARDENED_USERCOPY"), we need to reset pointer tags in
__check_heap_object() in mm/slab.c before doing any pointer math.

Link: http://lkml.kernel.org/r/9a5c0f958db10e69df5ff9f2b997866b56b7effc.1550602886.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Tested-by: Qian Cai <cai@lca.pw>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Evgeniy Stepanov <eugenis@google.com>
Cc: Kostya Serebryany <kcc@google.com>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 mm/slab.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/mm/slab.c b/mm/slab.c
index 9d5de959d9d9..05f21f736be8 100644
--- a/mm/slab.c
+++ b/mm/slab.c
@@ -4421,6 +4421,8 @@ void __check_heap_object(const void *ptr, unsigned long n, struct page *page,
 	unsigned int objnr;
 	unsigned long offset;
 
+	ptr = kasan_reset_tag(ptr);
+
 	/* Find and validate object. */
 	cachep = page->slab_cache;
 	objnr = obj_to_index(cachep, page, (void *)ptr);
-- 
2.19.1

[PATCH AUTOSEL 4.20 39/60] kasan, slab: make freelist stored without tags

[Sasha Levin]

From: Andrey Konovalov <andreyknvl@google.com>

[ Upstream commit 51dedad06b5f6c3eea7ec1069631b1ef7796912a ]

Similarly to "kasan, slub: move kasan_poison_slab hook before
page_address", move kasan_poison_slab() before alloc_slabmgmt(), which
calls page_address(), to make page_address() return value to be
non-tagged.  This, combined with calling kasan_reset_tag() for off-slab
slab management object, leads to freelist being stored non-tagged.

Link: http://lkml.kernel.org/r/dfb53b44a4d00de3879a05a9f04c1f55e584f7a1.1550602886.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Tested-by: Qian Cai <cai@lca.pw>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Evgeniy Stepanov <eugenis@google.com>
Cc: Kostya Serebryany <kcc@google.com>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 mm/slab.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/mm/slab.c b/mm/slab.c
index 05f21f736be8..b85524f2ab35 100644
--- a/mm/slab.c
+++ b/mm/slab.c
@@ -2381,6 +2381,7 @@ static void *alloc_slabmgmt(struct kmem_cache *cachep,
 		/* Slab management obj is off-slab. */
 		freelist = kmem_cache_alloc_node(cachep->freelist_cache,
 					      local_flags, nodeid);
+		freelist = kasan_reset_tag(freelist);
 		if (!freelist)
 			return NULL;
 	} else {
@@ -2694,6 +2695,13 @@ static struct page *cache_grow_begin(struct kmem_cache *cachep,
 
 	offset *= cachep->colour_off;
 
+	/*
+	 * Call kasan_poison_slab() before calling alloc_slabmgmt(), so
+	 * page_address() in the latter returns a non-tagged pointer,
+	 * as it should be for slab pages.
+	 */
+	kasan_poison_slab(page);
+
 	/* Get slab management. */
 	freelist = alloc_slabmgmt(cachep, page, offset,
 			local_flags & ~GFP_CONSTRAINT_MASK, page_node);
@@ -2702,7 +2710,6 @@ static struct page *cache_grow_begin(struct kmem_cache *cachep,
 
 	slab_map_pages(cachep, page, freelist);
 
-	kasan_poison_slab(page);
 	cache_init_objs(cachep, page);
 
 	if (gfpflags_allow_blocking(local_flags))
-- 
2.19.1