Re: Please pull 'upstream' branch of wireless-2.6

[Michael Buesch <mb@bu3sch.de>]

On Tuesday 12 September 2006 01:59, John W. Linville wrote:
> +	value16 = bcm43xx_shm_read16(bcm, BCM43xx_SHM_SHARED,
> +				     BCM43xx_UCODE_REVISION);
> +
> +	dprintk(KERN_INFO PFX "Microcode rev 0x%x, pl 0x%x "
> +		"(20%.2i-%.2i-%.2i  %.2i:%.2i:%.2i)\n", value16,
> +		bcm43xx_shm_read16(bcm, BCM43xx_SHM_SHARED,
> +				   BCM43xx_UCODE_PATCHLEVEL),
> +		(bcm43xx_shm_read16(bcm, BCM43xx_SHM_SHARED,
> +				    BCM43xx_UCODE_DATE) >> 12) & 0xf,
> +		(bcm43xx_shm_read16(bcm, BCM43xx_SHM_SHARED,
> +				    BCM43xx_UCODE_DATE) >> 8) & 0xf,
> +		bcm43xx_shm_read16(bcm, BCM43xx_SHM_SHARED,
> +				   BCM43xx_UCODE_DATE) & 0xff,
> +		(bcm43xx_shm_read16(bcm, BCM43xx_SHM_SHARED,
> +				   BCM43xx_UCODE_TIME) >> 11) & 0x1f,
> +		(bcm43xx_shm_read16(bcm, BCM43xx_SHM_SHARED,
> +				   BCM43xx_UCODE_TIME) >> 5) & 0x3f,
> +		bcm43xx_shm_read16(bcm, BCM43xx_SHM_SHARED,
> +				   BCM43xx_UCODE_TIME) & 0x1f);
> +
> +	if ( value16 > 0x128 ) {
> +		dprintk(KERN_ERR PFX
> +			"Firmware: no support for microcode rev > 0x128\n");
> +		err = -1;
> +		goto err_release_fw;
> +	}

Hm, this mustn't be a dprintk, as it's compiled away if
debugging is disabled. So it silently fails.
The text could be clarified, too.
returning -1 as error code is also very bad, as it's
propagated to userspace. I suggest EOPNOTSUPP, but maybe there's
something better.
Larry, can you do a patch which changes it to something like the following?

	if ( value16 > 0x128 ) {
		printk(KERN_ERR PFX
			"Firmware: no support for microcode extracted "
			"from version 4.x binary drivers.\n");
		err = -EOPNOTSUPP;
		goto err_release_fw;
	}

-- 
Greetings Michael.