From: Armin Wolf <W_Armin@gmx.de>
To: "Ilpo Järvinen" <ilpo.jarvinen@linux.intel.com>
Cc: linux@weissschuh.net, Hans de Goede <hdegoede@redhat.com>,
platform-driver-x86@vger.kernel.org,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] platform/x86: wmi-bmof: Make use of .bin_size() callback
Date: Fri, 13 Dec 2024 01:16:57 +0100 [thread overview]
Message-ID: <84bd3fe1-bfba-4427-8f01-7d1ab245e617@gmx.de> (raw)
In-Reply-To: <ad05e251-37bd-0cc2-8b11-a859e453f476@linux.intel.com>
Am 10.12.24 um 15:47 schrieb Ilpo Järvinen:
> On Fri, 6 Dec 2024, Armin Wolf wrote:
>
>> Until now the wmi-bmof driver had to allocate the binary sysfs
>> attribute dynamically since its size depends on the bmof buffer
>> returned by the firmware.
>>
>> Use the new .bin_size() callback to avoid having to do this memory
>> allocation.
>>
>> Tested on a Asus Prime B650-Plus.
>>
>> Signed-off-by: Armin Wolf <W_Armin@gmx.de>
>> ---
>> drivers/platform/x86/wmi-bmof.c | 75 +++++++++++++++++----------------
>> 1 file changed, 38 insertions(+), 37 deletions(-)
>>
>> diff --git a/drivers/platform/x86/wmi-bmof.c b/drivers/platform/x86/wmi-bmof.c
>> index df6f0ae6e6c7..3e33da36da8a 100644
>> --- a/drivers/platform/x86/wmi-bmof.c
>> +++ b/drivers/platform/x86/wmi-bmof.c
>> @@ -20,66 +20,66 @@
>>
>> #define WMI_BMOF_GUID "05901221-D566-11D1-B2F0-00A0C9062910"
>>
>> -struct bmof_priv {
>> - union acpi_object *bmofdata;
>> - struct bin_attribute bmof_bin_attr;
>> -};
>> -
>> -static ssize_t read_bmof(struct file *filp, struct kobject *kobj, struct bin_attribute *attr,
>> +static ssize_t bmof_read(struct file *filp, struct kobject *kobj, const struct bin_attribute *attr,
>> char *buf, loff_t off, size_t count)
>> {
>> - struct bmof_priv *priv = container_of(attr, struct bmof_priv, bmof_bin_attr);
>> + struct device *dev = kobj_to_dev(kobj);
>> + union acpi_object *obj = dev_get_drvdata(dev);
>>
>> - return memory_read_from_buffer(buf, count, &off, priv->bmofdata->buffer.pointer,
>> - priv->bmofdata->buffer.length);
>> + return memory_read_from_buffer(buf, count, &off, obj->buffer.pointer, obj->buffer.length);
>> }
>>
>> -static int wmi_bmof_probe(struct wmi_device *wdev, const void *context)
>> +static const BIN_ATTR_ADMIN_RO(bmof, 0);
>> +
>> +static const struct bin_attribute * const bmof_attrs[] = {
>> + &bin_attr_bmof,
>> + NULL
>> +};
>> +
>> +static size_t bmof_bin_size(struct kobject *kobj, const struct bin_attribute *attr, int n)
>> {
>> - struct bmof_priv *priv;
>> - int ret;
>> + struct device *dev = kobj_to_dev(kobj);
>> + union acpi_object *obj = dev_get_drvdata(dev);
>> +
>> + return obj->buffer.length;
>> +}
>>
>> - priv = devm_kzalloc(&wdev->dev, sizeof(struct bmof_priv), GFP_KERNEL);
>> - if (!priv)
>> - return -ENOMEM;
>> +static const struct attribute_group bmof_group = {
>> + .bin_size = bmof_bin_size,
>> + .bin_attrs_new = bmof_attrs,
>> +};
>> +
>> +static const struct attribute_group *bmof_groups[] = {
>> + &bmof_group,
>> + NULL
>> +};
>>
>> - dev_set_drvdata(&wdev->dev, priv);
>> +static int wmi_bmof_probe(struct wmi_device *wdev, const void *context)
>> +{
>> + union acpi_object *obj;
>>
>> - priv->bmofdata = wmidev_block_query(wdev, 0);
>> - if (!priv->bmofdata) {
>> + obj = wmidev_block_query(wdev, 0);
>> + if (!obj) {
>> dev_err(&wdev->dev, "failed to read Binary MOF\n");
>> return -EIO;
>> }
>>
>> - if (priv->bmofdata->type != ACPI_TYPE_BUFFER) {
>> + if (obj->type != ACPI_TYPE_BUFFER) {
>> dev_err(&wdev->dev, "Binary MOF is not a buffer\n");
>> - ret = -EIO;
>> - goto err_free;
>> + kfree(obj);
>> + return -EIO;
>> }
>>
>> - sysfs_bin_attr_init(&priv->bmof_bin_attr);
>> - priv->bmof_bin_attr.attr.name = "bmof";
>> - priv->bmof_bin_attr.attr.mode = 0400;
>> - priv->bmof_bin_attr.read = read_bmof;
>> - priv->bmof_bin_attr.size = priv->bmofdata->buffer.length;
>> -
>> - ret = device_create_bin_file(&wdev->dev, &priv->bmof_bin_attr);
>> - if (ret)
>> - goto err_free;
>> + dev_set_drvdata(&wdev->dev, obj);
>>
>> return 0;
>> -
>> - err_free:
>> - kfree(priv->bmofdata);
>> - return ret;
>> }
>>
>> static void wmi_bmof_remove(struct wmi_device *wdev)
>> {
>> - struct bmof_priv *priv = dev_get_drvdata(&wdev->dev);
>> + union acpi_object *obj = dev_get_drvdata(&wdev->dev);
>>
>> - device_remove_bin_file(&wdev->dev, &priv->bmof_bin_attr);
>> - kfree(priv->bmofdata);
>> + kfree(obj);
>> }
>>
>> static const struct wmi_device_id wmi_bmof_id_table[] = {
>> @@ -90,6 +90,7 @@ static const struct wmi_device_id wmi_bmof_id_table[] = {
>> static struct wmi_driver wmi_bmof_driver = {
>> .driver = {
>> .name = "wmi-bmof",
>> + .dev_groups = bmof_groups,
>> },
>> .probe = wmi_bmof_probe,
>> .remove = wmi_bmof_remove,
> So do I understand right that this meant to supercede the patch from
> Thomas?
>
Yes, i totally overlooked this patch, sorry. I will ask him to drop his patch.
Thanks,
Armin Wolf
next prev parent reply other threads:[~2024-12-13 0:17 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-12-06 21:56 Armin Wolf
2024-12-10 14:47 ` Ilpo Järvinen
2024-12-13 0:16 ` Armin Wolf [this message]
2024-12-13 11:28 ` Armin Wolf
2024-12-16 17:17 ` Ilpo Järvinen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=84bd3fe1-bfba-4427-8f01-7d1ab245e617@gmx.de \
--to=w_armin@gmx.de \
--cc=hdegoede@redhat.com \
--cc=ilpo.jarvinen@linux.intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux@weissschuh.net \
--cc=platform-driver-x86@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®