[PATCH 0/4] mm/madvise: cleanup requests validations and classifications

[PATCH 0/4] mm/madvise: cleanup requests validations and classifications

[SeongJae Park]

Cleanup madvise entry level code for cleaner request validations and
classifications.

Note that this series was a part of another one[1], and separated to
this one following a reviewer's nice comment[2].

Changes from the original series
(https://lore.kernel.org/20250310172318.653630-1-sj@kernel.org)
- Separate from the tlb flushes batching part
- Collect Reviewed-by: tags from Lorenzo
- Fix typos and wordsmith commit messages and comments
- Rebase on latest mm-unstable

[1] https://lore.kernel.org/20250310172318.653630-1-sj@kernel.org
[2] https://lore.kernel.org/0f90d56e-5960-4478-803e-1054696c0cde@lucifer.local

SeongJae Park (4):
  mm/madvise: use is_memory_failure() from madvise_do_behavior()
  mm/madvise: split out populate behavior check logic
  mm/madvise: deduplicate madvise_do_behavior() skip case handlings
  mm/madvise: remove len parameter of madvise_do_behavior()

 mm/madvise.c | 128 +++++++++++++++++++++++++++++----------------------
 1 file changed, 74 insertions(+), 54 deletions(-)


base-commit: 733c75c2264b9de72a463f731507e674a32ff094
-- 
2.39.5

[PATCH 1/4] mm/madvise: use is_memory_failure() from madvise_do_behavior()

[SeongJae Park]

To reduce redundant open-coded checks of CONFIG_MEMORY_FAILURE and
MADV_{HWPOISON,SOFT_OFFLINE} in madvise_[un]lock(), is_memory_failure()
is introduced.  madvise_do_behavior() is still doing the same open-coded
check, though.  Use is_memory_failure() instead.

To avoid build failure on !CONFIG_MEMORY_FAILURE case, implement an
empty madvise_inject_error() under the config.  Also move the definition
of is_memory_failure() inside #ifdef CONFIG_MEMORY_FAILURE clause for
madvise_inject_error() definition, to reduce duplicated ifdef clauses.

Signed-off-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
---
 mm/madvise.c | 49 +++++++++++++++++++++++++++----------------------
 1 file changed, 27 insertions(+), 22 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index 388dc289b5d1..c3ab1f283b18 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -1392,7 +1392,32 @@ static int madvise_inject_error(int behavior,
 
 	return 0;
 }
-#endif
+
+static bool is_memory_failure(int behavior)
+{
+	switch (behavior) {
+	case MADV_HWPOISON:
+	case MADV_SOFT_OFFLINE:
+		return true;
+	default:
+		return false;
+	}
+}
+
+#else
+
+static int madvise_inject_error(int behavior,
+		unsigned long start, unsigned long end)
+{
+	return 0;
+}
+
+static bool is_memory_failure(int behavior)
+{
+	return false;
+}
+
+#endif	/* CONFIG_MEMORY_FAILURE */
 
 static bool
 madvise_behavior_valid(int behavior)
@@ -1569,24 +1594,6 @@ int madvise_set_anon_name(struct mm_struct *mm, unsigned long start,
 }
 #endif /* CONFIG_ANON_VMA_NAME */
 
-#ifdef CONFIG_MEMORY_FAILURE
-static bool is_memory_failure(int behavior)
-{
-	switch (behavior) {
-	case MADV_HWPOISON:
-	case MADV_SOFT_OFFLINE:
-		return true;
-	default:
-		return false;
-	}
-}
-#else
-static bool is_memory_failure(int behavior)
-{
-	return false;
-}
-#endif
-
 static int madvise_lock(struct mm_struct *mm, int behavior)
 {
 	if (is_memory_failure(behavior))
@@ -1640,10 +1647,8 @@ static int madvise_do_behavior(struct mm_struct *mm,
 	unsigned long end;
 	int error;
 
-#ifdef CONFIG_MEMORY_FAILURE
-	if (behavior == MADV_HWPOISON || behavior == MADV_SOFT_OFFLINE)
+	if (is_memory_failure(behavior))
 		return madvise_inject_error(behavior, start, start + len_in);
-#endif
 	start = untagged_addr_remote(mm, start);
 	end = start + len;
 
-- 
2.39.5

Re: [PATCH 1/4] mm/madvise: use is_memory_failure() from madvise_do_behavior()

[Shakeel Butt]

On Wed, Mar 12, 2025 at 09:47:47AM -0700, SeongJae Park wrote:
> To reduce redundant open-coded checks of CONFIG_MEMORY_FAILURE and
> MADV_{HWPOISON,SOFT_OFFLINE} in madvise_[un]lock(), is_memory_failure()
> is introduced.  madvise_do_behavior() is still doing the same open-coded
> check, though.  Use is_memory_failure() instead.
> 
> To avoid build failure on !CONFIG_MEMORY_FAILURE case, implement an
> empty madvise_inject_error() under the config.  Also move the definition
> of is_memory_failure() inside #ifdef CONFIG_MEMORY_FAILURE clause for
> madvise_inject_error() definition, to reduce duplicated ifdef clauses.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>
> Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>

Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>

[PATCH 2/4] mm/madvise: split out populate behavior check logic

[SeongJae Park]

madvise_do_behavior() has a long open-coded 'behavior' check for
MADV_POPULATE_{READ,WRITE}.  It adds multiple layers[1] and make the
code arguably take longer time to read.  Like is_memory_failure(), split
out the check to a separate function.  This is not technically removing
the additional layer but discourage further extending the switch-case.
Also it makes madvise_do_behavior() code shorter and therefore easier to
read.

[1] https://lore.kernel.org/bd6d0bf1-c79e-46bd-a810-9791efb9ad73@lucifer.local

Signed-off-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
---
 mm/madvise.c | 20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index c3ab1f283b18..611db868ae38 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -1640,6 +1640,17 @@ static bool is_valid_madvise(unsigned long start, size_t len_in, int behavior)
 	return true;
 }
 
+static bool is_madvise_populate(int behavior)
+{
+	switch (behavior) {
+	case MADV_POPULATE_READ:
+	case MADV_POPULATE_WRITE:
+		return true;
+	default:
+		return false;
+	}
+}
+
 static int madvise_do_behavior(struct mm_struct *mm,
 		unsigned long start, size_t len_in, size_t len, int behavior)
 {
@@ -1653,16 +1664,11 @@ static int madvise_do_behavior(struct mm_struct *mm,
 	end = start + len;
 
 	blk_start_plug(&plug);
-	switch (behavior) {
-	case MADV_POPULATE_READ:
-	case MADV_POPULATE_WRITE:
+	if (is_madvise_populate(behavior))
 		error = madvise_populate(mm, start, end, behavior);
-		break;
-	default:
+	else
 		error = madvise_walk_vmas(mm, start, end, behavior,
 					  madvise_vma_behavior);
-		break;
-	}
 	blk_finish_plug(&plug);
 	return error;
 }
-- 
2.39.5

Re: [PATCH 2/4] mm/madvise: split out populate behavior check logic

[Shakeel Butt]

On Wed, Mar 12, 2025 at 09:47:48AM -0700, SeongJae Park wrote:
> madvise_do_behavior() has a long open-coded 'behavior' check for
> MADV_POPULATE_{READ,WRITE}.  It adds multiple layers[1] and make the
> code arguably take longer time to read.  Like is_memory_failure(), split
> out the check to a separate function.  This is not technically removing
> the additional layer but discourage further extending the switch-case.
> Also it makes madvise_do_behavior() code shorter and therefore easier to
> read.
> 
> [1] https://lore.kernel.org/bd6d0bf1-c79e-46bd-a810-9791efb9ad73@lucifer.local
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>
> Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>

Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>

[PATCH 3/4] mm/madvise: deduplicate madvise_do_behavior() skip case handlings

[SeongJae Park]

The logic for checking if a given madvise() request for a single memory
range can skip real work, namely madvise_do_behavior(), is duplicated in
do_madvise() and vector_madvise().  Split out the logic to a function
and reuse it.

Signed-off-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
---
 mm/madvise.c | 57 +++++++++++++++++++++++++++++++---------------------
 1 file changed, 34 insertions(+), 23 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index 611db868ae38..ba006d05c7ea 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -1640,6 +1640,31 @@ static bool is_valid_madvise(unsigned long start, size_t len_in, int behavior)
 	return true;
 }
 
+/*
+ * madvise_should_skip() - Return if the request is invalid or nothing.
+ * @start:	Start address of madvise-requested address range.
+ * @len_in:	Length of madvise-requested address range.
+ * @behavior:	Requested madvise behavor.
+ * @err:	Pointer to store an error code from the check.
+ *
+ * If the specified behaviour is invalid or nothing would occur, we skip the
+ * operation.  This function returns true in the cases, otherwise false.  In
+ * the former case we store an error on @err.
+ */
+static bool madvise_should_skip(unsigned long start, size_t len_in,
+		int behavior, int *err)
+{
+	if (!is_valid_madvise(start, len_in, behavior)) {
+		*err = -EINVAL;
+		return true;
+	}
+	if (start + PAGE_ALIGN(len_in) == start) {
+		*err = 0;
+		return true;
+	}
+	return false;
+}
+
 static bool is_madvise_populate(int behavior)
 {
 	switch (behavior) {
@@ -1747,23 +1772,15 @@ static int madvise_do_behavior(struct mm_struct *mm,
  */
 int do_madvise(struct mm_struct *mm, unsigned long start, size_t len_in, int behavior)
 {
-	unsigned long end;
 	int error;
-	size_t len;
-
-	if (!is_valid_madvise(start, len_in, behavior))
-		return -EINVAL;
-
-	len = PAGE_ALIGN(len_in);
-	end = start + len;
-
-	if (end == start)
-		return 0;
 
+	if (madvise_should_skip(start, len_in, behavior, &error))
+		return error;
 	error = madvise_lock(mm, behavior);
 	if (error)
 		return error;
-	error = madvise_do_behavior(mm, start, len_in, len, behavior);
+	error = madvise_do_behavior(mm, start, len_in, PAGE_ALIGN(len_in),
+			behavior);
 	madvise_unlock(mm, behavior);
 
 	return error;
@@ -1790,19 +1807,13 @@ static ssize_t vector_madvise(struct mm_struct *mm, struct iov_iter *iter,
 	while (iov_iter_count(iter)) {
 		unsigned long start = (unsigned long)iter_iov_addr(iter);
 		size_t len_in = iter_iov_len(iter);
-		size_t len;
-
-		if (!is_valid_madvise(start, len_in, behavior)) {
-			ret = -EINVAL;
-			break;
-		}
+		int error;
 
-		len = PAGE_ALIGN(len_in);
-		if (start + len == start)
-			ret = 0;
+		if (madvise_should_skip(start, len_in, behavior, &error))
+			ret = error;
 		else
-			ret = madvise_do_behavior(mm, start, len_in, len,
-					behavior);
+			ret = madvise_do_behavior(mm, start, len_in,
+					PAGE_ALIGN(len_in), behavior);
 		/*
 		 * An madvise operation is attempting to restart the syscall,
 		 * but we cannot proceed as it would not be correct to repeat
-- 
2.39.5

Re: [PATCH 3/4] mm/madvise: deduplicate madvise_do_behavior() skip case handlings

[Shakeel Butt]

On Wed, Mar 12, 2025 at 09:47:49AM -0700, SeongJae Park wrote:
> The logic for checking if a given madvise() request for a single memory
> range can skip real work, namely madvise_do_behavior(), is duplicated in
> do_madvise() and vector_madvise().  Split out the logic to a function
> and reuse it.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>
> Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>

Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>

[PATCH 4/4] mm/madvise: remove len parameter of madvise_do_behavior()

[SeongJae Park]

Because madise_should_skip() logic is factored out, making
madvise_do_behavior() calculates 'len' on its own rather then receiving
it as a parameter makes code simpler.  Remove the parameter.

Signed-off-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
---
 mm/madvise.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/mm/madvise.c b/mm/madvise.c
index ba006d05c7ea..b17f684322ad 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -1677,7 +1677,7 @@ static bool is_madvise_populate(int behavior)
 }
 
 static int madvise_do_behavior(struct mm_struct *mm,
-		unsigned long start, size_t len_in, size_t len, int behavior)
+		unsigned long start, size_t len_in, int behavior)
 {
 	struct blk_plug plug;
 	unsigned long end;
@@ -1686,7 +1686,7 @@ static int madvise_do_behavior(struct mm_struct *mm,
 	if (is_memory_failure(behavior))
 		return madvise_inject_error(behavior, start, start + len_in);
 	start = untagged_addr_remote(mm, start);
-	end = start + len;
+	end = start + PAGE_ALIGN(len_in);
 
 	blk_start_plug(&plug);
 	if (is_madvise_populate(behavior))
@@ -1779,8 +1779,7 @@ int do_madvise(struct mm_struct *mm, unsigned long start, size_t len_in, int beh
 	error = madvise_lock(mm, behavior);
 	if (error)
 		return error;
-	error = madvise_do_behavior(mm, start, len_in, PAGE_ALIGN(len_in),
-			behavior);
+	error = madvise_do_behavior(mm, start, len_in, behavior);
 	madvise_unlock(mm, behavior);
 
 	return error;
@@ -1812,8 +1811,7 @@ static ssize_t vector_madvise(struct mm_struct *mm, struct iov_iter *iter,
 		if (madvise_should_skip(start, len_in, behavior, &error))
 			ret = error;
 		else
-			ret = madvise_do_behavior(mm, start, len_in,
-					PAGE_ALIGN(len_in), behavior);
+			ret = madvise_do_behavior(mm, start, len_in, behavior);
 		/*
 		 * An madvise operation is attempting to restart the syscall,
 		 * but we cannot proceed as it would not be correct to repeat
-- 
2.39.5

Re: [PATCH 4/4] mm/madvise: remove len parameter of madvise_do_behavior()

[Shakeel Butt]

On Wed, Mar 12, 2025 at 09:47:50AM -0700, SeongJae Park wrote:
> Because madise_should_skip() logic is factored out, making
> madvise_do_behavior() calculates 'len' on its own rather then receiving
> it as a parameter makes code simpler.  Remove the parameter.
> 
> Signed-off-by: SeongJae Park <sj@kernel.org>
> Reviewed-by: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>

Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>

Re: [PATCH 0/4] mm/madvise: cleanup requests validations and classifications

[Lorenzo Stoakes]

On Wed, Mar 12, 2025 at 09:47:46AM -0700, SeongJae Park wrote:
> Cleanup madvise entry level code for cleaner request validations and
> classifications.
>
> Note that this series was a part of another one[1], and separated to
> this one following a reviewer's nice comment[2].

Thanks :>)

>
> Changes from the original series
> (https://lore.kernel.org/20250310172318.653630-1-sj@kernel.org)
> - Separate from the tlb flushes batching part
> - Collect Reviewed-by: tags from Lorenzo
> - Fix typos and wordsmith commit messages and comments
> - Rebase on latest mm-unstable
>
> [1] https://lore.kernel.org/20250310172318.653630-1-sj@kernel.org
> [2] https://lore.kernel.org/0f90d56e-5960-4478-803e-1054696c0cde@lucifer.local
>
> SeongJae Park (4):
>   mm/madvise: use is_memory_failure() from madvise_do_behavior()
>   mm/madvise: split out populate behavior check logic
>   mm/madvise: deduplicate madvise_do_behavior() skip case handlings
>   mm/madvise: remove len parameter of madvise_do_behavior()
>
>  mm/madvise.c | 128 +++++++++++++++++++++++++++++----------------------
>  1 file changed, 74 insertions(+), 54 deletions(-)
>
>
> base-commit: 733c75c2264b9de72a463f731507e674a32ff094
> --
> 2.39.5