Messages in this thread | | | Date | Wed, 21 May 2008 19:34:28 -0400 | From | "Mike Frysinger" <> | Subject | Re: + pcmcia-add-support-the-cf-pcmcia-driver-for-blackfin-try-2.patch added to -mm tree |
| |
On Tue, May 20, 2008 at 3:51 AM, Bryan Wu <cooloney@kernel.org> wrote: > On Tue, May 20, 2008 at 2:08 PM, <akpm@linux-foundation.org> wrote: >> +config BFIN_CFPCMCIA >> + tristate "Blackfin CompactFlash PCMCIA Driver" >> + depends on PCMCIA && BLACKFIN >> + help >> + Say Y here to support the CompactFlash PCMCIA driver for Blackfin.
please add a common statement that tells you the module name if the user builds it as a module.
>> + * author: Michael Hennerich (hennerich@blackfin.uclinux.org)
shouldnt this be MODULE_AUTHOR() ?
>> + u_int irq; >> + u_short cd_pfx;
u_int forms are weird ... i'd wonder why we have these variations anywhere in the kernel
>> +static int __init bfin_cf_probe(struct platform_device *pdev)
should be __devinit
>> + cd_pfx = platform_get_irq(pdev, 1); /*Card Detect GPIO PIN */ >> + if (cd_pfx > MAX_BLACKFIN_GPIOS) >> + return -EINVAL; >> + >> + if (gpio_request(cd_pfx, "pcmcia: CD")) {
i dont think that MAX_BLACKFIN_GPIOS check is needed. the gpio_request() function should already be doing it.
>> + printk(KERN_ERR >> + "BF5xx flash: Failed ro request Card Detect GPIO_%d\n",
"BF5xx flash" ?
>> + cf = kzalloc(sizeof *cf, GFP_KERNEL); >> + if (!cf) >> + return -ENOMEM;
we dont call gpio_free() here or in the fail0 case below. -mike
| |