[PATCH v2 0/2] kcov: improve mmap processing

[Aleksandr Nogikh]

Subsequent mmaps of the same kcov descriptor currently do not update the
virtual memory of the task and yet return 0 (success). This is
counter-intuitive and may lead to unexpected memory access errors.

Also, this unnecessarily limits the functionality of kcov to only the
simplest usage scenarios. Kcov instances are effectively forever attached
to their first address spaces and it becomes impossible to e.g. reuse the
same kcov handle in forked child processes without mmapping the memory
first. This is exactly what we tried to do in syzkaller and
inadvertently came upon this behavior.

This patch series addresses the problem described above.

v1 of the patch:
https://lore.kernel.org/lkml/20211220152153....@google.com/

Changes from v1:
- Split into 2 commits.
- Minor coding style changes.

Aleksandr Nogikh (2):
kcov: split ioctl handling into locked and unlocked parts
kcov: properly handle subsequent mmap calls

kernel/kcov.c | 99 +++++++++++++++++++++++++++++----------------------
1 file changed, 57 insertions(+), 42 deletions(-)

--
2.34.1.307.g9b7440fafd-goog

[Aleksandr Nogikh]

Currently all ioctls are de facto processed under a spin lock in order
to serialise them. This, however, prohibits the use of vmalloc and other
memory management functions in the implementation of those ioctls,
unnecessary complicating any further changes.

Let all ioctls first be processed inside the kcov_ioctl_unlocked()
function which should execute the ones that are not compatible with
spinlock and pass control to kcov_ioctl_locked() for all other ones.

Although it is still compatible with a spinlock, move KCOV_INIT_TRACE
handling to kcov_ioctl_unlocked(), so that its planned change is easier
to follow.

Signed-off-by: Aleksandr Nogikh <nog...@google.com>
---
kernel/kcov.c | 64 +++++++++++++++++++++++++++++++--------------------
1 file changed, 39 insertions(+), 25 deletions(-)

diff --git a/kernel/kcov.c b/kernel/kcov.c
index 36ca640c4f8e..5d87b4e0126f 100644
--- a/kernel/kcov.c
+++ b/kernel/kcov.c
@@ -564,31 +564,12 @@ static int kcov_ioctl_locked(struct kcov *kcov, unsigned int cmd,
unsigned long arg)
{
struct task_struct *t;
- unsigned long size, unused;
+ unsigned long flags, unused;
int mode, i;
struct kcov_remote_arg *remote_arg;
struct kcov_remote *remote;
- unsigned long flags;

switch (cmd) {
- case KCOV_INIT_TRACE:
- /*
- * Enable kcov in trace mode and setup buffer size.
- * Must happen before anything else.
- */
- if (kcov->mode != KCOV_MODE_DISABLED)
- return -EBUSY;
- /*
- * Size must be at least 2 to hold current position and one PC.
- * Later we allocate size * sizeof(unsigned long) memory,
- * that must not overflow.
- */
- size = arg;
- if (size < 2 || size > INT_MAX / sizeof(unsigned long))
- return -EINVAL;
- kcov->size = size;
- kcov->mode = KCOV_MODE_INIT;
- return 0;
case KCOV_ENABLE:
/*
* Enable coverage for the current task.
@@ -685,6 +666,43 @@ static int kcov_ioctl_locked(struct kcov *kcov, unsigned int cmd,
}
}

+static int kcov_ioctl_unlocked(struct kcov *kcov, unsigned int cmd,
+ unsigned long arg)
+{
+ unsigned long size, flags;
+ int res;
+
+ switch (cmd) {
+ case KCOV_INIT_TRACE:
+ /*
+ * Enable kcov in trace mode and setup buffer size.
+ * Must happen before anything else.
+ */
+ if (kcov->mode != KCOV_MODE_DISABLED)
+ return -EBUSY;
+ /*
+ * Size must be at least 2 to hold current position and one PC.
+ * Later we allocate size * sizeof(unsigned long) memory,
+ * that must not overflow.
+ */
+ size = arg;
+ if (size < 2 || size > INT_MAX / sizeof(unsigned long))
+ return -EINVAL;
+ kcov->size = size;
+ kcov->mode = KCOV_MODE_INIT;
+ return 0;
+ default:
+ /*
+ * All other commands can be fully executed under a spin lock, so we
+ * obtain and release it here to simplify the code of kcov_ioctl_locked().
+ */
+ spin_lock_irqsave(&kcov->lock, flags);
+ res = kcov_ioctl_locked(kcov, cmd, arg);
+ spin_unlock_irqrestore(&kcov->lock, flags);
+ return res;
+ }
+}
+
static long kcov_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
{
struct kcov *kcov;
@@ -692,7 +710,6 @@ static long kcov_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
struct kcov_remote_arg *remote_arg = NULL;
unsigned int remote_num_handles;
unsigned long remote_arg_size;
- unsigned long flags;

if (cmd == KCOV_REMOTE_ENABLE) {
if (get_user(remote_num_handles, (unsigned __user *)(arg +
@@ -713,10 +730,7 @@ static long kcov_ioctl(struct file *filep, unsigned int cmd, unsigned long arg)
}

kcov = filep->private_data;
- spin_lock_irqsave(&kcov->lock, flags);
- res = kcov_ioctl_locked(kcov, cmd, arg);
- spin_unlock_irqrestore(&kcov->lock, flags);
-
+ res = kcov_ioctl_unlocked(kcov, cmd, arg);
kfree(remote_arg);

return res;
--
2.34.1.307.g9b7440fafd-goog

[Aleksandr Nogikh]

Allocate the kcov buffer during KCOV_MODE_INIT in order to untie mmapping
of a kcov instance and the actual coverage collection process. Modify
kcov_mmap, so that it can be reliably used any number of times once
KCOV_MODE_INIT has succeeded.

These changes to the user-facing interface of the tool only weaken the
preconditions, so all existing user space code should remain compatible
with the new version.

Signed-off-by: Aleksandr Nogikh <nog...@google.com>
---

kernel/kcov.c | 49 +++++++++++++++++++++++++------------------------
1 file changed, 25 insertions(+), 24 deletions(-)

diff --git a/kernel/kcov.c b/kernel/kcov.c
index 5d87b4e0126f..d6a522fc6f36 100644
--- a/kernel/kcov.c
+++ b/kernel/kcov.c
@@ -459,37 +459,28 @@ void kcov_task_exit(struct task_struct *t)
static int kcov_mmap(struct file *filep, struct vm_area_struct *vma)
{
int res = 0;
- void *area;
struct kcov *kcov = vma->vm_file->private_data;
unsigned long size, off;
struct page *page;
unsigned long flags;

- area = vmalloc_user(vma->vm_end - vma->vm_start);
- if (!area)
- return -ENOMEM;
-
spin_lock_irqsave(&kcov->lock, flags);
size = kcov->size * sizeof(unsigned long);
- if (kcov->mode != KCOV_MODE_INIT || vma->vm_pgoff != 0 ||
+ if (kcov->area == NULL || vma->vm_pgoff != 0 ||
vma->vm_end - vma->vm_start != size) {
res = -EINVAL;
goto exit;
}
- if (!kcov->area) {
- kcov->area = area;
- vma->vm_flags |= VM_DONTEXPAND;
- spin_unlock_irqrestore(&kcov->lock, flags);
- for (off = 0; off < size; off += PAGE_SIZE) {
- page = vmalloc_to_page(kcov->area + off);
- if (vm_insert_page(vma, vma->vm_start + off, page))
- WARN_ONCE(1, "vm_insert_page() failed");
- }
- return 0;
+ spin_unlock_irqrestore(&kcov->lock, flags);
+ vma->vm_flags |= VM_DONTEXPAND;
+ for (off = 0; off < size; off += PAGE_SIZE) {
+ page = vmalloc_to_page(kcov->area + off);
+ if (vm_insert_page(vma, vma->vm_start + off, page))
+ WARN_ONCE(1, "vm_insert_page() failed");
}
+ return 0;
exit:
spin_unlock_irqrestore(&kcov->lock, flags);
- vfree(area);
return res;
}

@@ -671,25 +662,35 @@ static int kcov_ioctl_unlocked(struct kcov *kcov, unsigned int cmd,
{
unsigned long size, flags;
int res;
+ void *area;

switch (cmd) {
case KCOV_INIT_TRACE:
/*

* Enable kcov in trace mode and setup buffer size.

* Must happen before anything else.
- */
- if (kcov->mode != KCOV_MODE_DISABLED)
- return -EBUSY;
- /*
- * Size must be at least 2 to hold current position and one PC.
- * Later we allocate size * sizeof(unsigned long) memory,
- * that must not overflow.

+ *
+ * First check the size argument - it must be at least 2
+ * to hold the current position and one PC.
*/
size = arg;

if (size < 2 || size > INT_MAX / sizeof(unsigned long))

return -EINVAL;
+
+ area = vmalloc_user(size * sizeof(unsigned long));
+ if (area == NULL)
+ return -ENOMEM;
+
+ spin_lock_irqsave(&kcov->lock, flags);

+ if (kcov->mode != KCOV_MODE_DISABLED) {

+ spin_unlock_irqrestore(&kcov->lock, flags);
+ vfree(area);
+ return -EBUSY;
+ }
+ kcov->area = area;
kcov->size = size;
kcov->mode = KCOV_MODE_INIT;
+ spin_unlock_irqrestore(&kcov->lock, flags);
return 0;
default:
/*
--
2.34.1.307.g9b7440fafd-goog

[Marco Elver]

This patch should be a non-functional change, but it is not.

To do that, you'd have to add the locking around KCOV_INIT_TRACE here,
and then do whatever else you're doing in patch 2/2.

Also, I find that kcov_ioctl_unlocked() isn't a very descriptive name,
since now we have both locked and unlocked variants. What is it
actually doing?

Perhaps kcov_ioctl_with_context()? Assuming that 'struct kcov' is some
sort of context.

[Aleksandr Nogikh]

> To do that, you'd have to add the locking around KCOV_INIT_TRACE here,

Argh, indeed! Thanks, I'll fix it in v3.

> Also, I find that kcov_ioctl_unlocked() isn't a very descriptive name,
since now we have both locked and unlocked variants. What is it
actually doing?

The main motivation behind introducing that function was to get the
ability to do some processing outside of a spin lock without major
code refactoring. So it kind of wraps the existing ioctl processing
and, if it's KCOV_INIT_TRACE, performs the action itself.

I'm now thinking that we could probably do without introducing an
extra function at all - by moving `spin_lock_irqsave (&kcov->lock,
flags);` and `spin_unlock_irqrestore(&kcov->lock, flags);` into the
`kcov_ioctl_locked` function (and rename it into sth like
`kcov_do_ioctl`). So it could look like this:

switch (cmd) {
case KCOV_INIT_TRACE:
//...
}
spin_lock_irqsave (&kcov->lock, kcov_flags);
switch (cmd) {
case KCOV_ENABLE:
//...
default:
//...
}
spin_unlock_irqrestore(&kcov->lock, kcov_flags);