linux-kernel - Re: Grafting old platform drivers onto a new DT kernel

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: <5640D1F3.1010100@free.fr>
Date:	Mon, 9 Nov 2015 18:03:47 +0100
From:	Mason <slash.tmp@...e.fr>
To:	Mans Rullgard <mans@...sr.com>
Cc:	Javier Martinez Canillas <javier@...hile0.org>,
	Andrew Lunn <andrew@...n.ch>,
	LKML <linux-kernel@...r.kernel.org>,
	Linux ARM <linux-arm-kernel@...ts.infradead.org>,
	Marc Zyngier <marc.zyngier@....com>,
	Jason Cooper <jason@...edaemon.net>,
	Thomas Gleixner <tglx@...utronix.de>,
	Ulf Hansson <ulf.hansson@...aro.org>
Subject: Re: Grafting old platform drivers onto a new DT kernel

On 09/11/2015 17:12, Måns Rullgård wrote:

> Mason writes:
> 
>> On 09/11/2015 16:40, Måns Rullgård wrote:
>>
>>> The simplest solution for you is probably to add a quick and dirty DT
>>> binding to the old driver.  If it doesn't use any driver-specific
>>> platform data struct, you only need to set .of_match_table in the
>>> struct platform_driver.  If there is a platform data struct, you'll also
>>> need to write some code to populate it from DT properties.  It shouldn't
>>> take more than a few minutes per driver in most cases.
>>
>> I'll try that approach, although I fear that "a few minutes per driver"
>> is an optimistic assessment.
> 
> If the driver only needs an MMIO region and an IRQ, it is literally five
> lines of code.

It took me 7 days to figure out there were 2 lines missing in the
interrupt controller driver.

My problem is that I don't understand the platform API, nor the
interaction with the DT API.

Let me see...

In arch/arm/mach-tangox/platform_dev.c

static struct platform_device tangox_sdhci0_device = { ... };
static struct platform_device tangox_sdhci1_device = { ... };

static void tangox_init_sdhci(void)
{
	if (tangox_sdio_enabled(0))
		platform_device_register(&tangox_sdhci0_device);

	if (tangox_sdio_enabled(1))
		platform_device_register(&tangox_sdhci1_device);
}

called from tangox_init_devices() which is marked arch_initcall.



In the driver

static struct platform_driver tangox_platform_sdio0 = {
	.remove		= sdhci_tangox_remove,
	.suspend	= sdhci_tangox_suspend,
	.resume		= sdhci_tangox_resume,
	.driver		= {
		.name	= "tangox-sdhci",
		.owner	= THIS_MODULE,
	},
};


static struct platform_driver tangox_platform_sdio0 = {
	.remove		= sdhci_tangox_remove,
	.driver		= {
		.name	= "tangox-sdhci",
		.owner	= THIS_MODULE,
	},
};

static int __init tangox_sdhci_drv_init(void) {
	return platform_driver_probe(&tangox_platform_sdio0, sdhci_tangox_probe);
}

static void __exit tangox_sdhci_drv_exit(void) {
	platform_driver_unregister(&tangox_platform_sdio0);
}

module_init(tangox_sdhci_drv_init);
module_exit(tangox_sdhci_drv_exit);


The old way:

1) call platform_device_register() with a "struct platform_device"
2) call platform_driver_probe with a "struct platform_driver"

The new way(?)

The mess in 2) is hidden behind module_platform_driver?
The platform_device_register() is done by the DT core?
The struct platform_driver requires a probe function?

Regards.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/