lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <a40ef9a8-0589-4070-921e-f3461fa6759c@kernel.org>
Date: Tue, 5 Nov 2024 23:46:02 -0600
From: Mario Limonciello <superm1@...nel.org>
To: Armin Wolf <W_Armin@....de>, Mario Limonciello
 <mario.limonciello@....com>, Hans de Goede <hdegoede@...hat.com>,
 Ilpo Järvinen <ilpo.jarvinen@...ux.intel.com>
Cc: "Rafael J . Wysocki" <rafael@...nel.org>, Len Brown <lenb@...nel.org>,
 Maximilian Luz <luzmaximilian@...il.com>, Lee Chun-Yi <jlee@...e.com>,
 Shyam Sundar S K <Shyam-sundar.S-k@....com>,
 Corentin Chary <corentin.chary@...il.com>, "Luke D . Jones"
 <luke@...nes.dev>, Ike Panhc <ike.pan@...onical.com>,
 Henrique de Moraes Holschuh <hmh@....eng.br>,
 Alexis Belmonte <alexbelm48@...il.com>,
 Uwe Kleine-König <u.kleine-koenig@...gutronix.de>,
 Ai Chao <aichao@...inos.cn>, Gergo Koteles <soyer@....hu>,
 open list <linux-kernel@...r.kernel.org>,
 "open list:ACPI" <linux-acpi@...r.kernel.org>,
 "open list:MICROSOFT SURFACE PLATFORM PROFILE DRIVER"
 <platform-driver-x86@...r.kernel.org>,
 "open list:THINKPAD ACPI EXTRAS DRIVER"
 <ibm-acpi-devel@...ts.sourceforge.net>,
 Mark Pearson <mpearson-lenovo@...ebb.ca>,
 Matthew Schwartz <matthew.schwartz@...ux.dev>
Subject: Re: [PATCH v4 12/20] ACPI: platform_profile: Add profile attribute
 for class interface



On 11/5/24 22:10, Armin Wolf wrote:
> Am 05.11.24 um 16:33 schrieb Mario Limonciello:
> 
>> Reading and writing the `profile` sysfs file will use the callbacks for
>> the platform profile handler to read or set the given profile.
>>
>> Signed-off-by: Mario Limonciello <mario.limonciello@....com>
>> ---
>>   drivers/acpi/platform_profile.c | 118 ++++++++++++++++++++++++++++++++
>>   1 file changed, 118 insertions(+)
>>
>> diff --git a/drivers/acpi/platform_profile.c b/drivers/acpi/ 
>> platform_profile.c
>> index e1b6569c4ee70..79083d0bb22e3 100644
>> --- a/drivers/acpi/platform_profile.c
>> +++ b/drivers/acpi/platform_profile.c
>> @@ -65,6 +65,78 @@ static int _get_class_choices(struct device *dev, 
>> unsigned long *choices)
>>       return 0;
>>   }
>>
>> +/**
>> + * _store_class_profile - Set the profile for a class device
>> + * @dev: The class device
>> + * @data: The profile to set
>> + */
>> +static int _store_class_profile(struct device *dev, void *data)
>> +{
>> +    enum platform_profile_option profile;
>> +    unsigned long choices;
>> +    int *i = (int *)data;
>> +    int err;
>> +
>> +    err = _get_class_choices(dev, &choices);
>> +    if (err)
>> +        return err;
>> +
>> +    scoped_cond_guard(mutex_intr, return -ERESTARTSYS, &profile_lock) {
>> +        struct platform_profile_handler *handler;
>> +
>> +        if (!test_bit(*i, &choices))
>> +            return -EOPNOTSUPP;
>> +
>> +        handler = dev_get_drvdata(dev);
>> +        err = handler->profile_get(handler, &profile);
>> +        if (err)
>> +            return err;
>> +
>> +        err = handler->profile_set(handler, *i);
>> +        if (err) {
>> +            int recover_err;
>> +
>> +            dev_err(dev, "Failed to set profile: %d\n", err);
>> +            recover_err = handler->profile_set(handler, profile);
>> +            if (recover_err)
>> +                dev_err(dev, "Failed to reset profile: %d\n", 
>> recover_err);
>> +        }
> 
> The whole recovery handling seems unnecessary to me. In setting the 
> platform profile fails, then
> we should just return an error. The platform profile handler will tell 
> us the current platform
> profile anyway.

Sure, makes sense.  That also means no need to capture the profile 
before setting it.

> 
>> +        sysfs_notify(&handler->class_dev->kobj, NULL, 
>> "platform_profile");
>> +        kobject_uevent(&handler->class_dev->kobj, KOBJ_CHANGE);
> 
> Please avoid sending those events when the platform profile is changed 
> through the class sysfs interface.
> 
>> +    }
>> +
>> +    sysfs_notify(acpi_kobj, NULL, "platform_profile");
> 
> Please avoid sending this event when the platform profile is changed 
> through the legacy sysfs interface.

In both above cases - why?

* If I change using class interface then that implicitly means that 
legacy interface changes.
* If I change using legacy interface that implicitly means class 
interface changes too.

> 
>> +    return err ? err : 0;
>> +}
>> +
>> +/**
>> + * get_class_profile - Show the current profile for a class device
>> + * @dev: The class device
>> + * @profile: The profile to return
>> + * Return: 0 on success, -errno on failure
>> + */
>> +static int get_class_profile(struct device *dev,
>> +                 enum platform_profile_option *profile)
>> +{
>> +    struct platform_profile_handler *handler;
>> +    enum platform_profile_option val;
>> +    int err;
>> +
>> +    scoped_cond_guard(mutex_intr, return -ERESTARTSYS, &profile_lock) {
>> +        handler = dev_get_drvdata(dev);
>> +        err = handler->profile_get(handler, &val);
>> +        if (err) {
>> +            pr_err("Failed to get profile for handler %s\n", handler- 
>> >name);
>> +            return err;
>> +        }
>> +    }
>> +
>> +    if (WARN_ON(val >= PLATFORM_PROFILE_LAST))
>> +        return -EINVAL;
>> +    *profile = val;
>> +
>> +    return 0;
>> +}
>>
>>   /**
>>    * name_show - Show the name of the profile handler
>> @@ -102,12 +174,58 @@ static ssize_t choices_show(struct device *dev,
>>       return _commmon_choices_show(choices, buf);
>>   }
>>
>> +/**
>> + * profile_show - Show the current profile for a class device
>> + * @dev: The device
>> + * @attr: The attribute
>> + * @buf: The buffer to write to
>> + * Return: The number of bytes written
>> + */
>> +static ssize_t profile_show(struct device *dev,
>> +                struct device_attribute *attr,
>> +                char *buf)
>> +{
>> +    enum platform_profile_option profile = PLATFORM_PROFILE_LAST;
>> +    int err;
>> +
>> +    err = get_class_profile(dev, &profile);
>> +    if (err)
>> +        return err;
>> +
>> +    return sysfs_emit(buf, "%s\n", profile_names[profile]);
>> +}
>> +
>> +/**
>> + * profile_store - Set the profile for a class device
>> + * @dev: The device
>> + * @attr: The attribute
>> + * @buf: The buffer to read from
>> + * @count: The number of bytes to read
>> + * Return: The number of bytes read
>> + */
>> +static ssize_t profile_store(struct device *dev,
>> +                 struct device_attribute *attr,
>> +                 const char *buf, size_t count)
>> +{
>> +    int i, ret;
>> +
>> +    i = sysfs_match_string(profile_names, buf);
>> +    if (i < 0)
>> +        return -EINVAL;
>> +
>> +    ret = _store_class_profile(dev, (void *)(long)&i);
> 
> Please just pass &i.

Ack.

> 
> Thanks,
> Armin Wolf
> 
>> +
>> +    return ret ? ret : count;
>> +}
>>
>>   static DEVICE_ATTR_RO(name);
>>   static DEVICE_ATTR_RO(choices);
>> +static DEVICE_ATTR_RW(profile);
>> +
>>   static struct attribute *profile_attrs[] = {
>>       &dev_attr_name.attr,
>>       &dev_attr_choices.attr,
>> +    &dev_attr_profile.attr,
>>       NULL
>>   };
>>   ATTRIBUTE_GROUPS(profile);
> 


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ