[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHp75VcZqpQ7qM2aBzSsqa+xYOdevZkWy=h+u8=nhdP2foyH4g@mail.gmail.com>
Date: Fri, 1 Jun 2018 12:54:04 +0300
From: Andy Shevchenko <andy.shevchenko@...il.com>
To: Bastian Germann <bastiangermann@...hpost.de>
Cc: Guenter Roeck <linux@...ck-us.net>,
Luca Tettamanti <kronos.it@...il.com>,
Jean Delvare <jdelvare@...e.com>, linux-hwmon@...r.kernel.org,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/2] hwmon: (asus_atk0110) Make use of device managed memory
On Fri, Jun 1, 2018 at 1:57 AM, Bastian Germann
<bastiangermann@...hpost.de> wrote:
> Use devm_* variants of kstrdup and kzalloc. Get rid of the kfree cleanups.
> static int atk_debugfs_ggrp_open(struct inode *inode, struct file *file)
> {
> struct atk_data *data = inode->i_private;
> + struct device *dev = &data->acpi_dev->dev;
> char *buf = NULL;
> union acpi_object *ret;
> u8 cls;
> @@ -748,7 +748,7 @@ static int atk_debugfs_ggrp_open(struct inode *inode, struct file *file)
> id = &pack->package.elements[0];
> if (id->integer.value == data->debugfs.id) {
> /* Print the package */
> - buf = kzalloc(512, GFP_KERNEL);
> + buf = devm_kzalloc(dev, 512, GFP_KERNEL);
> if (!buf) {
> ACPI_FREE(ret);
> return -ENOMEM;
Looking to the function name, it feels like you are creating a memory
leak and devm_ is inappropriate here.
> @@ -776,16 +776,9 @@ static ssize_t atk_debugfs_ggrp_read(struct file *file, char __user *buf,
> return simple_read_from_buffer(buf, count, pos, str, len);
> }
>
> -static int atk_debugfs_ggrp_release(struct inode *inode, struct file *file)
> -{
> - kfree(file->private_data);
> - return 0;
> -}
> -
> static const struct file_operations atk_debugfs_ggrp_fops = {
> .read = atk_debugfs_ggrp_read,
> .open = atk_debugfs_ggrp_open,
> - .release = atk_debugfs_ggrp_release,
> .llseek = no_llseek,
> };
So do these.
Also the question of time to leave of the objects used for debugfs
communication. Please double check that it's not affected by the
change.
Otherwise looks good.
--
With Best Regards,
Andy Shevchenko
Powered by blists - more mailing lists