[bpf-next,v4,1/5] bpf: Introduce task_file open-coded iterator kfuncs

Commit Message

Juntong Deng Nov. 19, 2024, 5:53 p.m. UTC

This patch adds the open-coded iterator style process file iterator
kfuncs bpf_iter_task_file_{new,next,destroy} that iterates over all
files opened by the specified process.

bpf_iter_task_file_next returns a pointer to bpf_iter_task_file_item,
which currently contains *task, *file, fd. This is an extensible
structure that enables compatibility with different versions
through CO-RE.

The reference to struct file acquired by the previous
bpf_iter_task_file_next() is released in the next
bpf_iter_task_file_next(), and the last reference is released in the
last bpf_iter_task_file_next() that returns NULL.

In the bpf_iter_task_file_destroy(), if the iterator does not iterate to
the end, then the last struct file reference is released at this time.

Signed-off-by: Juntong Deng <[e-mail]>
---
 kernel/bpf/helpers.c   |  3 ++
 kernel/bpf/task_iter.c | 92 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 95 insertions(+)

Comments

Jiri Olsa Nov. 20, 2024, 10:55 a.m. UTC | #1

On Tue, Nov 19, 2024 at 05:53:58PM +0000, Juntong Deng wrote:

SNIP

> +/**
> + * bpf_iter_task_file_next() - Get the next file in bpf_iter_task_file
> + *
> + * bpf_iter_task_file_next acquires a reference to the struct file.
> + *
> + * The reference to struct file acquired by the previous
> + * bpf_iter_task_file_next() is released in the next bpf_iter_task_file_next(),
> + * and the last reference is released in the last bpf_iter_task_file_next()
> + * that returns NULL.
> + *
> + * @it: the bpf_iter_task_file to be checked
> + *
> + * @returns a pointer to bpf_iter_task_file_item
> + */
> +__bpf_kfunc struct bpf_iter_task_file_item *bpf_iter_task_file_next(struct bpf_iter_task_file *it)
> +{
> +	struct bpf_iter_task_file_kern *kit = (void *)it;
> +	struct bpf_iter_task_file_item *item = &kit->item;
> +
> +	if (item->file)
> +		fput(item->file);
> +

missing rcu_read_lock ?

jirka

> +	item->file = task_lookup_next_fdget_rcu(item->task, &kit->next_fd);
> +	item->fd = kit->next_fd;
> +
> +	kit->next_fd++;
> +
> +	if (!item->file)
> +		return NULL;
> +
> +	return item;
> +}
> +
> +/**
> + * bpf_iter_task_file_destroy() - Destroy a bpf_iter_task_file
> + *
> + * If the iterator does not iterate to the end, then the last
> + * struct file reference is released at this time.
> + *
> + * @it: the bpf_iter_task_file to be destroyed
> + */
> +__bpf_kfunc void bpf_iter_task_file_destroy(struct bpf_iter_task_file *it)
> +{
> +	struct bpf_iter_task_file_kern *kit = (void *)it;
> +	struct bpf_iter_task_file_item *item = &kit->item;
> +
> +	if (item->file)
> +		fput(item->file);
> +}
> +
>  __bpf_kfunc_end_defs();
>  
>  DEFINE_PER_CPU(struct mmap_unlock_irq_work, mmap_unlock_work);
> -- 
> 2.39.5
>

Jiri Olsa Nov. 20, 2024, 11:04 a.m. UTC | #2

On Wed, Nov 20, 2024 at 11:55:23AM +0100, Jiri Olsa wrote:
> On Tue, Nov 19, 2024 at 05:53:58PM +0000, Juntong Deng wrote:
> 
> SNIP
> 
> > +/**
> > + * bpf_iter_task_file_next() - Get the next file in bpf_iter_task_file
> > + *
> > + * bpf_iter_task_file_next acquires a reference to the struct file.
> > + *
> > + * The reference to struct file acquired by the previous
> > + * bpf_iter_task_file_next() is released in the next bpf_iter_task_file_next(),
> > + * and the last reference is released in the last bpf_iter_task_file_next()
> > + * that returns NULL.
> > + *
> > + * @it: the bpf_iter_task_file to be checked
> > + *
> > + * @returns a pointer to bpf_iter_task_file_item
> > + */
> > +__bpf_kfunc struct bpf_iter_task_file_item *bpf_iter_task_file_next(struct bpf_iter_task_file *it)
> > +{
> > +	struct bpf_iter_task_file_kern *kit = (void *)it;
> > +	struct bpf_iter_task_file_item *item = &kit->item;
> > +
> > +	if (item->file)
> > +		fput(item->file);
> > +
> 
> missing rcu_read_lock ?

nah user needs to take it explicitly, should have read the whole thing first, sry

jirka

> 
> jirka
> 
> > +	item->file = task_lookup_next_fdget_rcu(item->task, &kit->next_fd);
> > +	item->fd = kit->next_fd;
> > +
> > +	kit->next_fd++;
> > +
> > +	if (!item->file)
> > +		return NULL;
> > +
> > +	return item;
> > +}
> > +
> > +/**
> > + * bpf_iter_task_file_destroy() - Destroy a bpf_iter_task_file
> > + *
> > + * If the iterator does not iterate to the end, then the last
> > + * struct file reference is released at this time.
> > + *
> > + * @it: the bpf_iter_task_file to be destroyed
> > + */
> > +__bpf_kfunc void bpf_iter_task_file_destroy(struct bpf_iter_task_file *it)
> > +{
> > +	struct bpf_iter_task_file_kern *kit = (void *)it;
> > +	struct bpf_iter_task_file_item *item = &kit->item;
> > +
> > +	if (item->file)
> > +		fput(item->file);
> > +}
> > +
> >  __bpf_kfunc_end_defs();
> >  
> >  DEFINE_PER_CPU(struct mmap_unlock_irq_work, mmap_unlock_work);
> > -- 
> > 2.39.5
> >

Patch

diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index 751c150f9e1c..ed99d2bfd425 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -3130,6 +3130,9 @@  BTF_ID_FLAGS(func, bpf_iter_css_destroy, KF_ITER_DESTROY)
 BTF_ID_FLAGS(func, bpf_iter_task_new, KF_ITER_NEW | KF_TRUSTED_ARGS | KF_RCU_PROTECTED)
 BTF_ID_FLAGS(func, bpf_iter_task_next, KF_ITER_NEXT | KF_RET_NULL)
 BTF_ID_FLAGS(func, bpf_iter_task_destroy, KF_ITER_DESTROY)
+BTF_ID_FLAGS(func, bpf_iter_task_file_new, KF_ITER_NEW | KF_TRUSTED_ARGS | KF_RCU_PROTECTED)
+BTF_ID_FLAGS(func, bpf_iter_task_file_next, KF_ITER_NEXT | KF_RET_NULL)
+BTF_ID_FLAGS(func, bpf_iter_task_file_destroy, KF_ITER_DESTROY)
 BTF_ID_FLAGS(func, bpf_dynptr_adjust)
 BTF_ID_FLAGS(func, bpf_dynptr_is_null)
 BTF_ID_FLAGS(func, bpf_dynptr_is_rdonly)
diff --git a/kernel/bpf/task_iter.c b/kernel/bpf/task_iter.c
index 5af9e130e500..7a3d8c9c3480 100644
--- a/kernel/bpf/task_iter.c
+++ b/kernel/bpf/task_iter.c
@@ -1031,6 +1031,98 @@  __bpf_kfunc void bpf_iter_task_destroy(struct bpf_iter_task *it)
 {
 }
 
+struct bpf_iter_task_file_item {
+	struct task_struct *task;
+	struct file *file;
+	unsigned int fd;
+} __aligned(8);
+
+struct bpf_iter_task_file {
+	__u64 __opaque[4];
+} __aligned(8);
+
+struct bpf_iter_task_file_kern {
+	struct bpf_iter_task_file_item item;
+	unsigned int next_fd;
+} __aligned(8);
+
+/**
+ * bpf_iter_task_file_new() - Initialize a new task file iterator for a task,
+ * used to iterate over all files opened by a specified task
+ *
+ * Note that task file iterator requires to be used within RCU CS.
+ *
+ * @it: the new bpf_iter_task_file to be created
+ * @task: a pointer pointing to a task to be iterated over
+ */
+__bpf_kfunc int bpf_iter_task_file_new(struct bpf_iter_task_file *it,
+		struct task_struct *task)
+{
+	struct bpf_iter_task_file_kern *kit = (void *)it;
+	struct bpf_iter_task_file_item *item = &kit->item;
+
+	BUILD_BUG_ON(sizeof(struct bpf_iter_task_file_kern) > sizeof(struct bpf_iter_task_file));
+	BUILD_BUG_ON(__alignof__(struct bpf_iter_task_file_kern) !=
+		     __alignof__(struct bpf_iter_task_file));
+
+	item->task = task;
+	item->file = NULL;
+	item->fd = 0;
+	kit->next_fd = 0;
+
+	return 0;
+}
+
+/**
+ * bpf_iter_task_file_next() - Get the next file in bpf_iter_task_file
+ *
+ * bpf_iter_task_file_next acquires a reference to the struct file.
+ *
+ * The reference to struct file acquired by the previous
+ * bpf_iter_task_file_next() is released in the next bpf_iter_task_file_next(),
+ * and the last reference is released in the last bpf_iter_task_file_next()
+ * that returns NULL.
+ *
+ * @it: the bpf_iter_task_file to be checked
+ *
+ * @returns a pointer to bpf_iter_task_file_item
+ */
+__bpf_kfunc struct bpf_iter_task_file_item *bpf_iter_task_file_next(struct bpf_iter_task_file *it)
+{
+	struct bpf_iter_task_file_kern *kit = (void *)it;
+	struct bpf_iter_task_file_item *item = &kit->item;
+
+	if (item->file)
+		fput(item->file);
+
+	item->file = task_lookup_next_fdget_rcu(item->task, &kit->next_fd);
+	item->fd = kit->next_fd;
+
+	kit->next_fd++;
+
+	if (!item->file)
+		return NULL;
+
+	return item;
+}
+
+/**
+ * bpf_iter_task_file_destroy() - Destroy a bpf_iter_task_file
+ *
+ * If the iterator does not iterate to the end, then the last
+ * struct file reference is released at this time.
+ *
+ * @it: the bpf_iter_task_file to be destroyed
+ */
+__bpf_kfunc void bpf_iter_task_file_destroy(struct bpf_iter_task_file *it)
+{
+	struct bpf_iter_task_file_kern *kit = (void *)it;
+	struct bpf_iter_task_file_item *item = &kit->item;
+
+	if (item->file)
+		fput(item->file);
+}
+
 __bpf_kfunc_end_defs();
 
 DEFINE_PER_CPU(struct mmap_unlock_irq_work, mmap_unlock_work);