When the buffer attributes were wrapped in iio_dev_attr types, I forgot to
duplicate the names, so that when iio_free_chan_devattr_list() gets called
on cleanup, these get free'd.
I stumbled over this while accidentally breaking a driver doing
iio_device_register(), and then the issue appeared.

Some ways to fix this are:
1. Just use kstrdup() during iio_buffer_wrap_attr()
2. Just use kfree_const() during iio_free_chan_devattr_list
3. Use both kstrdup_const() & kfree_const() (in the places mentioned above)

This implements the third option, as it allows some users/drivers to
specify some attributes allocated on the heap.

Fixes: a1a11142f66c ("iio: buffer: wrap all buffer attributes into 
iio_dev_attr")
Signed-off-by: Alexandru Ardelean <alexandru.ardel...@analog.com>
---
 drivers/iio/industrialio-buffer.c | 1 +
 drivers/iio/industrialio-core.c   | 2 +-
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/iio/industrialio-buffer.c 
b/drivers/iio/industrialio-buffer.c
index 5d641f8adfbd..ac882e60c419 100644
--- a/drivers/iio/industrialio-buffer.c
+++ b/drivers/iio/industrialio-buffer.c
@@ -1306,6 +1306,7 @@ static struct attribute *iio_buffer_wrap_attr(struct 
iio_buffer *buffer,
                return NULL;
 
        iio_attr->buffer = buffer;
+       iio_attr->dev_attr.attr.name = kstrdup_const(attr->name, GFP_KERNEL);
        memcpy(&iio_attr->dev_attr, dattr, sizeof(iio_attr->dev_attr));
 
        list_add(&iio_attr->l, &buffer->buffer_attr_list);
diff --git a/drivers/iio/industrialio-core.c b/drivers/iio/industrialio-core.c
index 0d8c6e88d993..cb2735d2ae4b 100644
--- a/drivers/iio/industrialio-core.c
+++ b/drivers/iio/industrialio-core.c
@@ -1358,7 +1358,7 @@ void iio_free_chan_devattr_list(struct list_head 
*attr_list)
        struct iio_dev_attr *p, *n;
 
        list_for_each_entry_safe(p, n, attr_list, l) {
-               kfree(p->dev_attr.attr.name);
+               kfree_const(p->dev_attr.attr.name);
                list_del(&p->l);
                kfree(p);
        }
-- 
2.25.1

Reply via email to