Re: [PATCH v2 5/7] [PATCH 5/8] can: SJA1000 generic platform bus driver

[Wolfgang Grandegger <wg@grandegger.com>]

Jonathan Corbet wrote:
> On Tue, 12 May 2009 11:28:02 +0200
> Wolfgang Grandegger <wg@grandegger.com> wrote:
> 
>> This driver adds support for the SJA1000 chips connected to the
>> "platform bus", which can be found on various embedded systems.
> 
> [...]
> 
>> +
>> +static u8 sp_read_reg(const struct net_device *dev, int reg)
>> +{
>> +	return ioread8((void __iomem *)(dev->base_addr + reg));
>> +}
>> +
>> +static void sp_write_reg(const struct net_device *dev, int reg, u8 val)
>> +{
>> +	iowrite8(val, (void __iomem *)(dev->base_addr + reg));
>> +}
> 
> So there's no locking around accesses to the hardware at all.  How do you
> protect against concurrent access?

There is no concurrent access to the same register and PCI register
accesses do not need to be serialized.

> [...]
> 
>> +static int sp_remove(struct platform_device *pdev)
>> +{
>> +	struct net_device *dev = dev_get_drvdata(&pdev->dev);
>> +	struct resource *res;
>> +
>> +	unregister_sja1000dev(dev);
>> +	dev_set_drvdata(&pdev->dev, NULL);
>> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> +	release_mem_region(res->start, res->end - res->start + 1);
>> +
>> +	if (dev->base_addr)
>> +		iounmap((void __iomem *)dev->base_addr);
> 
> Seems like you should unmap it before releasing it back to the kernel.
> Nobody else is ever going to jump in and try to map it, but still...

Will fix.

Wolfgang.