[PATCH v6 2/4] power: reset: add reboot mode driver
Krzysztof Kozlowski
k.kozlowski at samsung.com
Sun Mar 27 23:34:32 PDT 2016
On 24.03.2016 17:03, Andy Yan wrote:
> Hi Krzystof:
(...)
>>> +static int get_reboot_mode_magic(struct reboot_mode_driver *reboot,
>>> + const char *cmd)
>>> +{
>>> + const char *normal = "normal";
>>> + int magic = 0;
>>> + struct mode_info *info;
>>> +
>>> + if (!cmd)
>>> + cmd = normal;
>>> +
>>> + list_for_each_entry(info, &reboot->head, list) {
>>> + if (!strcmp(info->mode, cmd)) {
>>> + magic = info->magic;
>>> + break;
>>> + }
>>> + }
>>> +
>>> + return magic;
>> In absence of 'normal' mode (it is not described as required property)
>> the magic will be '0'. It would be nice to document that in bindings.
>> Imagine someone forgets about this and will wonder why 0x0 is written
>> to his precious register on normal reboot...
>
> If the magic value is '0', we won't touch the register, please see
> reboot_mode_notify bellow.
Ah, indeed... so we cannot use value of '0' for magic (e.g. to clear any
existing value for normal reboot)?
>> It would be nice to document that 'mode-normal' has a special
>> (hard-coded) meaning.
>>
>>> +}
>>> +
>>> +static int reboot_mode_notify(struct notifier_block *this,
>>> + unsigned long mode, void *cmd)
>>> +{
>>> + struct reboot_mode_driver *reboot;
>>> + int magic;
>>> +
>>> + reboot = container_of(this, struct reboot_mode_driver,
>>> reboot_notifier);
>>> + magic = get_reboot_mode_magic(reboot, cmd);
>>> + if (magic)
>>> + reboot->write(magic);
>>> +
>>> + return NOTIFY_DONE;
>>> +}
>>> +
>>> +int reboot_mode_register(struct device *dev, int (*write)(int))
>>> +{
>>> + struct reboot_mode_driver *reboot;
>>> + struct mode_info *info;
>>> + struct property *prop;
>>> + size_t len = strlen(PREFIX);
>>> + int ret;
>>> +
>>> + reboot = devm_kzalloc(dev, sizeof(*reboot), GFP_KERNEL);
>>> + if (!reboot)
>>> + return -ENOMEM;
>>> +
>>> + reboot->write = write;
>>> + INIT_LIST_HEAD(&reboot->head);
>>> + for_each_property_of_node(dev->of_node, prop) {
>>> + if (len > strlen(prop->name) || strncmp(prop->name,
>>> PREFIX, len))
>>> + continue;
>> New line please for readability.
>
> Okay
>>
>>> + info = devm_kzalloc(dev, sizeof(*info), GFP_KERNEL);
>>> + if (!info)
>>> + return -ENOMEM;
>> Ditto.
>
> Okay
>>
>>> + strcpy(info->mode, prop->name + len);
>> Ehm, and how do you protect that name of mode is shorter than 32
>> characters?
>
> How about info->mode = prop->name + len ?
I don't get your answer.
As fair as I read the code, the prop->name can be very long and you are
copying it from 5 character. If the name of the mode has bazillion
characters then again my question: how do you protect that it will fit
in 32 bytes of 'mode'?
>>
>>> + if (of_property_read_u32(dev->of_node, prop->name,
>>> &info->magic)) {
>>> + dev_err(dev, "reboot mode %s without magic
>>> number\n",
>>> + info->mode);
>>> + devm_kfree(dev, info);
>>> + continue;
>>> + }
>>> + list_add_tail(&info->list, &reboot->head);
>>> + }
>>> + reboot->reboot_notifier.notifier_call = reboot_mode_notify;
>>> + ret = register_reboot_notifier(&reboot->reboot_notifier);
>>> + if (ret)
>>> + dev_err(dev, "can't register reboot notifier\n");
>>> +
>>> + return ret;
>>> +}
>>> +EXPORT_SYMBOL_GPL(reboot_mode_register);
>>> +
>>> +MODULE_AUTHOR("Andy Yan <andy.yan at rock-chips.com");
>>> +MODULE_DESCRIPTION("System reboot mode driver");
>>> +MODULE_LICENSE("GPL v2");
>>> diff --git a/drivers/power/reset/reboot-mode.h
>>> b/drivers/power/reset/reboot-mode.h
>>> new file mode 100644
>>> index 0000000..44ed34f
>>> --- /dev/null
>>> +++ b/drivers/power/reset/reboot-mode.h
>>> @@ -0,0 +1,6 @@
>>> +#ifndef __REBOOT_MODE_H__
>>> +#define __REBOOT_MODE_H__
>>> +
>>> +int reboot_mode_register(struct device *dev, int (*write)(int));
>> Documentation would be appreciated. Although it is local header but
>> you decoupled them and you are exporting the function.
>
> I think I should remove the EXPORT_SYMBOL_GPL.
You need to export symbols which are located in modules. Is it located
in module?
Best regards,
Krzysztof
More information about the Linux-rockchip
mailing list