Re: [PATCH] iio: buffer: Fix memory leak in iio_buffers_alloc_sysfs_and_mask()

[Jonathan Cameron <jic23@kernel.org>]

On Wed, 13 Oct 2021 22:41:20 +0800
Yang Yingliang <yangyingliang@huawei.com> wrote:

> When 'iio_dev_opaque->buffer_ioctl_handler' alloc fails in
> iio_buffers_alloc_sysfs_and_mask(), the 'attrs' allocated in
> iio_buffer_register_legacy_sysfs_groups() will be leaked:
> 
> unreferenced object 0xffff888108568d00 (size 128):
>   comm "88", pid 2014, jiffies 4294963294 (age 26.920s)
>   hex dump (first 32 bytes):
>     80 3e da 02 80 88 ff ff 00 3a da 02 80 88 ff ff  .>.......:......
>     00 35 da 02 80 88 ff ff 00 38 da 02 80 88 ff ff  .5.......8......
>   backtrace:
>     [<0000000095a9e51e>] __kmalloc+0x1a3/0x2f0
>     [<00000000faa3735e>] iio_buffers_alloc_sysfs_and_mask+0xfa3/0x1480 [industrialio]
>     [<00000000a46384dc>] __iio_device_register+0x52e/0x1b40 [industrialio]
>     [<00000000210af05e>] __devm_iio_device_register+0x22/0x80 [industrialio]
>     [<00000000730d7b41>] adjd_s311_probe+0x195/0x200 [adjd_s311]
>     [<00000000c0f70eb9>] i2c_device_probe+0xa07/0xbb0
>     [<00000000b721e9ec>] really_probe+0x285/0xc30
>     [<00000000e6690642>] __driver_probe_device+0x35f/0x4f0
>     [<000000003ee4152f>] driver_probe_device+0x4f/0x140
>     [<0000000055cfdd2f>] __device_attach_driver+0x24c/0x330
>     [<0000000098033282>] bus_for_each_drv+0x15d/0x1e0
>     [<00000000fa1876f5>] __device_attach+0x267/0x410
>     [<00000000838ca724>] bus_probe_device+0x1ec/0x2a0
>     [<00000000f4ce978e>] device_add+0xc3d/0x2020
>     [<000000000ee12e9b>] i2c_new_client_device+0x614/0xb00
>     [<00000000fc5d3221>] new_device_store+0x1f4/0x410
> 
> Call iio_buffer_unregister_legacy_sysfs_groups() in error path
> to fix this.
> 
> Reported-by: Hulk Robot <hulkci@huawei.com>
> Fixes: d9a625744ed0 ("iio: core: merge buffer/ & scan_elements/ attributes")
> Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Hi

Another good find, but I'd like to handle this one slightly differently.

This is directly calling the unwind of something done deep in a function called
by this one.  That's not a clear structure though I can see you copied it from
iio_buffers_free_sysfs_and_mask() (which will also need modification).

This cleanup needs to be in __iio_buffer_free_sysfs_and_mask() which will need to
take an additional parameter (index) so that we know we are on index == 0 and hence
do the legacy cleanup.

Thanks,

Jonathan


> ---
>  drivers/iio/industrialio-buffer.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/iio/industrialio-buffer.c b/drivers/iio/industrialio-buffer.c
> index ae0912a14578..ba1c5c898e53 100644
> --- a/drivers/iio/industrialio-buffer.c
> +++ b/drivers/iio/industrialio-buffer.c
> @@ -1630,7 +1630,7 @@ int iio_buffers_alloc_sysfs_and_mask(struct iio_dev *indio_dev)
>  	iio_dev_opaque->buffer_ioctl_handler = kzalloc(sz, GFP_KERNEL);
>  	if (!iio_dev_opaque->buffer_ioctl_handler) {
>  		ret = -ENOMEM;
> -		goto error_unwind_sysfs_and_mask;
> +		goto error_unregister_legacy_sysfs_groups;
>  	}
>  
>  	iio_dev_opaque->buffer_ioctl_handler->ioctl = iio_device_buffer_ioctl;
> @@ -1639,6 +1639,8 @@ int iio_buffers_alloc_sysfs_and_mask(struct iio_dev *indio_dev)
>  
>  	return 0;
>  
> +error_unregister_legacy_sysfs_groups:
> +	iio_buffer_unregister_legacy_sysfs_groups(indio_dev);
>  error_unwind_sysfs_and_mask:
>  	for (; unwind_idx >= 0; unwind_idx--) {
>  		buffer = iio_dev_opaque->attached_buffers[unwind_idx];